Import packages¶
In [ ]:
import mne
import numpy as np
import scipy
from scipy.signal import savgol_filter
from scipy.stats import trim_mean
from sklearn.manifold import MDS, TSNE
from sklearn.cluster import KMeans
from sklearn.covariance import shrunk_covariance
import copy
import torch
import seaborn as sns
import matplotlib.pyplot as plt
from matplotlib.pyplot import MultipleLocator
from pylab import mpl
import seaborn as sns
import dill
import warnings
warnings.filterwarnings('ignore')
from utils import UDEC_Network, draw_states, ttest_for_clusters
In [ ]:
plt.rcParams['font.family']=['Arial', 'Times New Roman']
plt.style.use('default')
mpl.rcParams["axes.unicode_minus"] = False
%config InlineBackend.figure_format = 'svg'
font = {'family':['Arial', 'Times New Roman'], 'color':'k', 'weight':'normal', 'size':10 }
colors = sns.color_palette('tab10')
device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
print(device)
mne.cuda.init_cuda(verbose=True)
cuda:0 Now using CUDA device 0 Enabling CUDA with 10.09 GB available memory
Load EEG data¶
In [ ]:
subject_num = 40
erp_data = np.zeros((subject_num, 2, 28, 256)) # [subject, type, ch, time]
trim = lambda x: trim_mean(x, 0.2, axis=0)
file_path_pre = './N400_Data/'
file_path_post = '_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set'
for sub in range(subject_num):
if (sub+1) in [1,3,9,16,23,25,27,28,29,40]: #[3,16,23,28,40]: #
continue
file_name = file_path_pre + str(sub+1) + file_path_post
# from -200 ms to 800 ms, baseline correction performed
epo = mne.io.read_epochs_eeglab(file_name);
# epo.ch_names
eog_ch_list = ['HEOG_left', 'HEOG_right', 'VEOG_lower', '(corr) HEOG',
'(corr) VEOG', '(uncorr) HEOG', '(uncorr) VEOG']
epo.drop_channels(eog_ch_list);
reject = dict(eeg=100e-6 ) # unit: V (EEG channels)
flat_criteria = dict(eeg=1e-6)
epo.drop_bad(reject, flat=flat_criteria, verbose=False);
# epo.plot_drop_log();
# epo.plot(events=True);
events_dict = { }
for word in ['unrela', 'rela']:
events_dict[word] = []
for id in epo.event_id:
# Bin 1 Unrelated
# Bin 2 Related
if id[1] == '1':
events_dict['unrela'].append(id)
elif id[1] == '2':
events_dict['rela'].append(id)
for i, word in enumerate(['unrela', 'rela']):
erp_data[sub, i] = epo[events_dict[word]].average(method=trim).get_data()
Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\2_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 108 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\4_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 113 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\5_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 113 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\6_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 111 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\7_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 117 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\8_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 115 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\10_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 108 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\11_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 114 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\12_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 113 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\13_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 112 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\14_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 110 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\15_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 107 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\17_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 114 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\18_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 114 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\19_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 119 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\20_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 109 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\21_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 116 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\22_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 112 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\24_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 115 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\26_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 118 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\30_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 118 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\31_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 117 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\32_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 109 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\33_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 117 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\34_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 116 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\35_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 105 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\36_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 112 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\37_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 117 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\38_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 114 matching events found No baseline correction applied 0 projection items activated Ready. Extracting parameters from e:\论文写作\论文2数据分析\N400 - 副本\N400_Data\39_N400_shifted_ds_reref_ucbip_hpfilt_ica_corr_cbip_elist_bins_epoch_interp_ar.set... Not setting metadata 118 matching events found No baseline correction applied 0 projection items activated Ready.
In [ ]:
erp_data_mean = np.mean(erp_data, axis=0) #[2, 28, 256]
erp_data_norm = erp_data_mean * 1e6
sub_num = erp_data.shape[0]
type_num = erp_data_mean.shape[0]
ch_num = erp_data_mean.shape[1]
time_num = erp_data_mean.shape[2]
print(erp_data_mean.shape)
(2, 28, 256)
Plot ERP¶
In [ ]:
info_tmp = mne.create_info(ch_names=epo.ch_names, sfreq=256, ch_types='eeg')
erp1 = mne.EvokedArray(erp_data_mean[0], info_tmp, tmin=-0.2, nave=None)
erp1.set_montage(epo.get_montage());
erp2 = mne.EvokedArray(erp_data_mean[1], info_tmp, tmin=-0.2, nave=None)
erp2.set_montage(epo.get_montage());
erp_diff = mne.EvokedArray(erp_data_mean[0]-erp_data_mean[1], info_tmp, tmin=-0.2, nave=None)
erp_diff.set_montage(epo.get_montage());
In [ ]:
fig = erp1.plot(gfp=True, ylim = dict(eeg=[-8, 10]));
fig = erp2.plot(gfp=True, ylim = dict(eeg=[-8, 15]));
fig = erp_diff.plot(gfp=True, ylim = dict(eeg=[-15, 5]));
In [ ]:
tp_args = { "vlim":(-10, 2), "time_format":'%3.2f', "contours":10, "cmap":"jet" }
erp_diff.plot_joint(title='Unrelated minus Related', times=[0.25, 0.34, 0.45, 0.55], topomap_args=tp_args);
No projector specified for this dataset. Please consider the method self.add_proj.
In [ ]:
erp_diff.plot_topomap(cmap="jet", vlim=[-10, 2], time_format='%3.2f',
times=[0.2, 0.25,0.3,0.35,0.4,0.45,0.5,0.55,0.6,0.65], nrows=1 );
In [ ]:
ch_index = epo.ch_names.index('CPz')
erp_data_1 = erp_data_mean[0][ch_index] * 1e6
erp_data_2 = erp_data_mean[1][ch_index] * 1e6
n_sub = erp_data.shape[0] # [40, 2, 28, 256]
erp_data_diff_sem = np.std((erp_data[:,0,ch_index]-erp_data[:,1,ch_index]) * 1e6, axis=0, ddof=0)/np.sqrt(n_sub)
colors = sns.color_palette('Set2')
plt.figure(figsize=(3.5, 2))
ax = plt.gca()
# ax.axvspan(0, 100, alpha=0.1, color='grey')
plt.grid(color='gray', linewidth=0.5, alpha=0.5, linestyle='-')
plt.axhline(0.0, color='k', linewidth=1.0, linestyle=':', alpha=0.5)
plt.axvline(51.2, color='k', linewidth=1.0, linestyle=':', alpha=0.5)
# filter
mean_line1 = savgol_filter(erp_data_1, 4, 2) # win_size: 4, order: 3
mean_line2 = savgol_filter(erp_data_2, 4, 2) # win_size: 4, order: 3
diff_line = mean_line1 - mean_line2
sem_line1 = diff_line - erp_data_diff_sem
sem_line2 = diff_line + erp_data_diff_sem
plt.plot(mean_line1, label='Unrelated', color=colors[0])
plt.plot(mean_line2, label='Related', color=colors[1])
plt.plot(diff_line, label='Unrelated minus Related', linewidth=1.0, color=colors[2])
plt.fill_between(range(256), sem_line1, sem_line2, alpha=0.5, label='SEM', color=colors[2] )
ax.xaxis.set_minor_locator(MultipleLocator(12.8))
plt.xlim([0,256])
plt.ylim([-12, 12])
plt.legend(prop={'family':'Times New Roman', 'size':9}, ncol = 1, bbox_to_anchor=(1.0, 1.0))
font = {'family' : 'Times New Roman', # 'Microsoft YaHei' 'SimHei' 'serif'
'color' : 'k',
'weight' : 'normal',
'size' : 10,
}
plt.xticks(np.linspace(0, 256, 11, endpoint=True), ['-0.2', '', '0.0', '', '0.2', '', '0.4', '', '0.6', '', '0.8'], fontdict=font);
# ['-0.2', '-0.1', '0.0', '0.1', '0.2', '0.3', '0.4', '0.5', '0.6', '0.7', '0.8']
plt.xlabel("Time (s)", fontdict=font)
plt.yticks(np.linspace(-12, 12, 5, endpoint=True), np.linspace(-12, 12, 5, endpoint=True), fontdict=font);
plt.ylabel("CPz (uV)", fontdict=font);
Calculate spatial covariance matrix¶
In [ ]:
# time [-0.2, 0.8] 256Hz
# epo.time_as_index(0.0) 51
# epo.time_as_index(0.7) 231
# slide_window 0.1s [-12, 12]
half_win = 12
start_t = 51
end_t = 231
len_t = 180
cov_mat = 0
cov_diag = 0
cov_data = 0
crop_erp = 0
flag = 0
for tp in range(type_num):
data = erp_data_norm[tp]
cov_mat_epoch = np.empty((len_t, ch_num, ch_num))
cov_data_epoch = np.empty(( len_t, int(ch_num*((ch_num+1)/2)) ))
cov_diag_epoch = np.empty(( len_t, ch_num ))
for ind, t in enumerate(range(start_t, end_t)):
# calculate covariance matrix
cov = np.cov(data[:, t-half_win:t+half_win])
# get the triangle elements of the matrix
cov_data_epoch[ind] = cov[np.triu_indices(cov.shape[0], k=0)]
cov_diag_epoch[ind] = np.diagonal(cov)
# shrunk
cov = shrunk_covariance(cov, shrinkage=0.01)
cov_mat_epoch[ind] = cov
if flag == 0:
flag = 1
cov_mat = cov_mat_epoch
cov_diag = cov_diag_epoch
cov_data = cov_data_epoch
crop_erp = erp_data_mean[tp, :, start_t:end_t].T
else:
cov_data = np.vstack((cov_data, cov_data_epoch))
cov_diag = np.vstack((cov_diag, cov_diag_epoch))
cov_mat = np.vstack((cov_mat, cov_mat_epoch))
crop_erp = np.vstack((crop_erp, erp_data_mean[tp, :, start_t:end_t].T))
print( crop_erp.shape ) # [times, chs]
print( cov_mat.shape ) # [times, chs, chs]
print( cov_diag.shape ) # [times, chs]
print( cov_data.shape ) # [times, diag]
print( type_num*len_t, int(ch_num*((ch_num+1)/2)) ) # 360 = 2 * 180
(360, 28) (360, 28, 28) (360, 28) (360, 406) 360 406
In [ ]:
with open('./tmp_data/raw_eeg_data.pkl', 'wb') as f:
dill.dump([erp_data, erp_data_mean, cov_mat, cov_data, cov_diag, crop_erp], f)
# with open('./tmp_data/raw_eeg_data.pkl', 'rb') as f:
# [erp_data, erp_data_mean, cov_mat, cov_data, cov_diag, crop_erp] = dill.load(f)
Plot covariance matrices¶
In [ ]:
def draw_cov_mat(ind):
fig, ax = plt.subplots(figsize=(1.5, 1.5))
ax0 = ax.matshow(cov_mat[ind], interpolation='none', vmin=-2, vmax=2.0, cmap='jet');# cmap='turbo',
clb = fig.colorbar(ax0, fraction=0.045);
ax.xaxis.set_ticks_position("bottom")
plt.grid(color='gray', linestyle=':', linewidth=0.5);
clb.set_ticks(ticks=[-2, 0, 2])
cbar_label = clb.ax.get_xticklabels() + clb.ax.get_yticklabels()
[lab.set_font('Times New Roman') for lab in cbar_label] #Times New Roman Arial
clb.ax.tick_params(labelsize=10)
ax.set_xticks(np.arange(0,30,10), np.arange(0,30,10), fontdict=font);
ax.set_yticks(np.arange(0,30,10), np.arange(0,30,10), fontdict=font);
plt.show();
tps = [20, 50, 80]
for tp in tps:
draw_cov_mat(tp)
Calculate spatial pattern distance¶
distance in observation space¶
In [ ]:
%%writefile calc_dist_o_fun.py
import scipy
import numpy as np
def calc_dist_o(cov1, cov2):
evals, _ = scipy.linalg.eigh(cov1,cov2)
res = evals[-1] #np.max(evals)
return res
Overwriting calc_dist_o_fun.py
In [ ]:
import calc_dist_o_fun
import multiprocessing
num_cores = int(multiprocessing.cpu_count())
print("CPU cores: ", num_cores)
CPU cores: 32
In [ ]:
tp = cov_mat.shape[0]
cov_dist_o = np.full((tp, tp), np.nan)
for i in range(tp):
if i % 200 == 0: # ~3min
print(i)
pool = multiprocessing.Pool(processes = num_cores)
res = []
for j in range(i+1, tp):
r = pool.apply_async(calc_dist_o_fun.calc_dist_o, args=( cov_mat[i], cov_mat[j] ))
res.append(r)
pool.close()
evals_i = np.squeeze( [p.get() for p in res] )
cov_dist_o[i, i+1:] = evals_i
0 200
In [ ]:
# with open('./tmp_data/cov_dist_o.pkl', 'wb') as f:
# dill.dump(cov_dist_o, f)
with open('./tmp_data/cov_dist_o.pkl', 'rb') as f:
cov_dist_o = dill.load(f)
In [ ]:
fig, ax = plt.subplots(figsize=(1.7, 1.7))
ax0 = ax.matshow(np.log(cov_dist_o), interpolation='none', vmin=1.0, vmax=10.0, cmap='jet');
clb = fig.colorbar(ax0, fraction=0.045);
ax.xaxis.set_ticks_position("bottom")
plt.grid(color='gray', linestyle=':', linewidth=0.5);
clb.set_ticks(ticks=[1, 5, 10])
cbar_label = clb.ax.get_xticklabels() + clb.ax.get_yticklabels()
[lab.set_font('Arial') for lab in cbar_label] #Times New Roman #set_fontstyle
clb.ax.tick_params(labelsize=10)
ax.set_xticks(np.arange(0,400,100), np.arange(0,400,100), fontdict=font);
ax.set_yticks(np.arange(0,400,100), np.arange(0,400,100), fontdict=font);
plt.show();
In [ ]:
norm_cov_dist_o = np.log(cov_dist_o)
norm_cov_dist_o /= (np.nanmax(norm_cov_dist_o))
In [ ]:
fig, ax = plt.subplots(figsize=(1.7, 1.7))
ax0 = ax.matshow(norm_cov_dist_o, interpolation='none', vmin=0.0, vmax=1.0, cmap='jet', );
clb = fig.colorbar(ax0, fraction=0.045);
ax.xaxis.set_ticks_position("bottom")
plt.grid(color='gray', linestyle=':', linewidth=0.5);
clb.set_ticks(ticks=[0, 0.5, 1])
cbar_label = clb.ax.get_xticklabels() + clb.ax.get_yticklabels()
[lab.set_font('Arial') for lab in cbar_label]
clb.ax.tick_params(labelsize=10)
ax.set_xticks(np.arange(0,400,100), np.arange(0,400,100), fontdict=font);
ax.set_yticks(np.arange(0,400,100), np.arange(0,400,100), fontdict=font);
plt.show();
distance in source space¶
In [ ]:
%%writefile calc_dist_s_fun.py
import scipy
import numpy as np
def calc_dist_s(cov1, cov2):
d2 = (np.log(scipy.linalg.eigvalsh(cov1,cov2))**2).sum(axis=-1)
riemann_dist = np.sqrt(d2)
return riemann_dist
Overwriting calc_dist_s_fun.py
In [ ]:
import calc_dist_s_fun
import multiprocessing
num_cores = int(multiprocessing.cpu_count())
print("CPU cores: ", num_cores)
CPU cores: 32
In [ ]:
tp = cov_mat.shape[0]
cov_dist_s = np.full((tp, tp), np.nan)
for i in range(tp):
if i % 200 == 0: # ~3min
print(i)
pool = multiprocessing.Pool(processes = num_cores)
res = []
for j in range(i+1, tp):
r = pool.apply_async(calc_dist_s_fun.calc_dist_s, args=( cov_mat[i], cov_mat[j] ))
res.append(r)
pool.close()
riemann_dist_i = np.squeeze( [p.get() for p in res] )
cov_dist_s[i, i+1:] = riemann_dist_i
0 200
In [ ]:
# with open('./tmp_data/cov_dist_s.pkl', 'wb') as f:
# dill.dump(cov_dist_s, f)
with open('./tmp_data/cov_dist_s.pkl', 'rb') as f:
cov_dist_s = dill.load(f)
In [ ]:
fig, ax = plt.subplots(figsize=(1.7, 1.7))
ax0 = ax.matshow(cov_dist_s, interpolation='none', vmin=0.0, vmax=20.0, cmap='jet', );#
clb = fig.colorbar(ax0, fraction=0.045);
ax.xaxis.set_ticks_position("bottom")
plt.grid(color='gray', linestyle=':', linewidth=0.5);
clb.set_ticks(ticks=[0, 10, 20])
cbar_label = clb.ax.get_xticklabels() + clb.ax.get_yticklabels()
[lab.set_font('Arial') for lab in cbar_label] #Times New Roman #set_fontstyle
clb.ax.tick_params(labelsize=10)
ax.set_xticks(np.arange(0,400,100), np.arange(0,400,100), fontdict=font);
ax.set_yticks(np.arange(0,400,100), np.arange(0,400,100), fontdict=font);
plt.show();
In [ ]:
norm_cov_dist_s = copy.deepcopy(cov_dist_s)
norm_cov_dist_s /= (np.nanmax(norm_cov_dist_s))
In [ ]:
fig, ax = plt.subplots(figsize=(1.7, 1.7))
ax0 = ax.matshow(norm_cov_dist_s, interpolation='none', vmin=0.0, vmax=1.0, cmap='jet', );
clb = fig.colorbar(ax0, fraction=0.045);
ax.xaxis.set_ticks_position("bottom")
plt.grid(color='gray', linestyle=':', linewidth=0.5);
clb.set_ticks(ticks=[0, 0.5, 1])
cbar_label = clb.ax.get_xticklabels() + clb.ax.get_yticklabels()
[lab.set_font('Arial') for lab in cbar_label]
clb.ax.tick_params(labelsize=10)
ax.set_xticks(np.arange(0,400,100), np.arange(0,400,100), fontdict=font);
ax.set_yticks(np.arange(0,400,100), np.arange(0,400,100), fontdict=font);
plt.show();
Multi-dimensional Scaling¶
In [ ]:
norm_dist_o_symm = np.ones_like(norm_cov_dist_o)
for i in range(norm_cov_dist_o.shape[0]):
for j in range(i+1, norm_cov_dist_o.shape[0]):
norm_dist_o_symm[i,j] = norm_cov_dist_o[i,j]
norm_dist_o_symm[j,i] = norm_cov_dist_o[i,j]
mds = MDS(n_components=28, dissimilarity='precomputed', metric=True,
n_jobs=32, random_state=3, normalized_stress='auto')
features_dist_o = mds.fit_transform(norm_dist_o_symm)
features_dist_o.shape
Out[ ]:
(360, 28)
In [ ]:
norm_dist_s_symm = np.ones_like(norm_cov_dist_s)
for i in range(norm_cov_dist_s.shape[0]):
for j in range(i+1, norm_cov_dist_s.shape[0]):
norm_dist_s_symm[i,j] = norm_cov_dist_s[i,j]
norm_dist_s_symm[j,i] = norm_cov_dist_s[i,j]
mds = MDS(n_components=28, dissimilarity='precomputed', metric=True,
n_jobs=32, random_state=3, normalized_stress='auto')
features_dist_s = mds.fit_transform(norm_dist_s_symm)
features_dist_s.shape
Out[ ]:
(360, 28)
In [ ]:
# with open('./tmp_data/distance_mds_data.pkl', 'wb') as f:
# dill.dump([features_dist_o, features_dist_s], f)
with open('./tmp_data/distance_mds_data.pkl', 'rb') as f:
[features_dist_o, features_dist_s] = dill.load(f)
Data normlization¶
In [ ]:
norm_cov_diag = np.log(10*cov_diag)
norm_dist_o = 20*features_dist_o
norm_dist_s = 20*features_dist_s
In [ ]:
import warnings
warnings.filterwarnings('ignore')
font = {'family':['Times New Roman', 'Arial', ], 'weight':'normal', 'size':10 }
colors = sns.color_palette('Set2')
plt.figure(figsize=(2.5, 1.5))
sns.histplot(norm_cov_diag.flatten(), bins=80, stat='density', legend=False, color=colors[0], alpha=0.8, )
sns.histplot(norm_dist_o.flatten()+0.2, bins=80, stat='density', legend=False, color=colors[1], alpha=0.8, )
sns.histplot(norm_dist_s.flatten()-0.2, bins=80, stat='density', legend=False, color=colors[2], alpha=0.8, )
plt.xlim([-6, 6])
plt.ylim([0, 0.4])
plt.legend(labels=['Diagonal elements', 'Distance o', 'Distance s'], prop=font, bbox_to_anchor=(1.0, 1.0));
In [ ]:
input_data_raw = np.hstack((norm_cov_diag, norm_dist_o, norm_dist_s))
print(input_data_raw.shape) #[360, 84]
input_data = torch.from_numpy(input_data_raw).type(torch.FloatTensor) #[samples, features]
device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
print(device)
(360, 84) cuda:0
In [ ]:
tsne = TSNE(n_components=2, init='pca', random_state=0, n_jobs=-1, perplexity=30)
for data in [norm_cov_diag, norm_dist_o, norm_dist_s, input_data_raw]:
tsne_2d = tsne.fit_transform(data)
fig, axi1=plt.subplots(1, figsize=(2, 1.5))
axi1.scatter(tsne_2d[:, 0], tsne_2d[:, 1],
marker='*', s=10, color=sns.color_palette('Paired')[1],
)
ax = plt.gca()
plt.grid(True, linewidth=0.5, color='gray', linestyle=':')
plt.xlim([-30, 30])
plt.ylim([-30, 30])
# plt.yticks([-100,-50,0,50,100],[-100,-50,0,50,100])
# plt.xticks([-100,-50,0,50,100],[-100,-50,0,50,100])
ax.tick_params(which='both', bottom=True, top=False, left=True, right=False,
labelbottom=True, labelleft=True, direction='out',width=1)
plt.show()
In [ ]:
# with open('./tmp_data/network_input_data.pkl', 'wb') as f:
# dill.dump([norm_cov_diag, norm_dist_o, norm_dist_s, input_data_raw], f)
with open('./tmp_data/network_input_data.pkl', 'rb') as f:
[norm_cov_diag, norm_dist_o, norm_dist_s, input_data_raw] = dill.load(f)
Train AutoEncoder Network¶
In [ ]:
# import importlib
# importlib.reload(UDEC_Network)
EPOCHS_PRE = 3001
LR_PRE = 5e-3 # dynamic adjust learning rate
BATCH_SIZE = 32
autoencoder = UDEC_Network.AutoEncoder().to(device)
checkpoint = { "epoch": 0, "best": float("inf") }
file_path_prefix = './network_data/'
ae_save_path = file_path_prefix + 'autoencoder.pth'
UDEC_Network.pretrain(data=input_data, model=autoencoder, savepath=ae_save_path,
checkpoint=checkpoint, file_path_prefix=file_path_prefix,
num_epochs=EPOCHS_PRE, batch_size=BATCH_SIZE, lr=LR_PRE)
epoch [1/3001], MSE_loss:1.66254 epoch [2/3001], MSE_loss:1.13898 epoch [3/3001], MSE_loss:1.01787 epoch [4/3001], MSE_loss:0.96542 epoch [5/3001], MSE_loss:0.83371 epoch [6/3001], MSE_loss:1.01575 epoch [7/3001], MSE_loss:0.92843 epoch [8/3001], MSE_loss:0.80495 epoch [9/3001], MSE_loss:0.81596 epoch [10/3001], MSE_loss:0.77381 epoch [11/3001], MSE_loss:0.74888 epoch [12/3001], MSE_loss:0.58345 epoch [13/3001], MSE_loss:0.74370 epoch [14/3001], MSE_loss:0.50986 epoch [15/3001], MSE_loss:0.61735 epoch [16/3001], MSE_loss:0.48200 epoch [17/3001], MSE_loss:0.49468 epoch [18/3001], MSE_loss:0.51220 epoch [19/3001], MSE_loss:0.41671 epoch [20/3001], MSE_loss:0.28173 epoch [21/3001], MSE_loss:0.30446 epoch [22/3001], MSE_loss:0.36370 epoch [23/3001], MSE_loss:0.30576 epoch [24/3001], MSE_loss:0.44834 epoch [25/3001], MSE_loss:0.24842 epoch [26/3001], MSE_loss:0.43508 epoch [27/3001], MSE_loss:0.55455 epoch [28/3001], MSE_loss:0.45840 epoch [29/3001], MSE_loss:0.23072 epoch [30/3001], MSE_loss:0.21870 epoch [31/3001], MSE_loss:0.26832 epoch [32/3001], MSE_loss:0.29121 epoch [33/3001], MSE_loss:0.28388 epoch [34/3001], MSE_loss:0.24028 epoch [35/3001], MSE_loss:0.26064 epoch [36/3001], MSE_loss:0.26895 epoch [37/3001], MSE_loss:0.20207 epoch [38/3001], MSE_loss:0.15183 epoch [39/3001], MSE_loss:0.27656 epoch [40/3001], MSE_loss:0.21840 epoch [41/3001], MSE_loss:0.20838 epoch [42/3001], MSE_loss:0.15333 epoch [43/3001], MSE_loss:0.16620 epoch [44/3001], MSE_loss:0.12631 epoch [45/3001], MSE_loss:0.16376 epoch [46/3001], MSE_loss:0.25585 epoch [47/3001], MSE_loss:0.19916 epoch [48/3001], MSE_loss:0.24977 epoch [49/3001], MSE_loss:0.30111 epoch [50/3001], MSE_loss:0.16175 epoch [51/3001], MSE_loss:0.30097 epoch [52/3001], MSE_loss:0.16872 epoch [53/3001], MSE_loss:0.19043 epoch [54/3001], MSE_loss:0.15995 epoch [55/3001], MSE_loss:0.20287 epoch [56/3001], MSE_loss:0.23826 epoch [57/3001], MSE_loss:0.24307 epoch [58/3001], MSE_loss:0.17195 epoch [59/3001], MSE_loss:0.13154 epoch [60/3001], MSE_loss:0.12628 epoch [61/3001], MSE_loss:0.18054 epoch [62/3001], MSE_loss:0.15759 epoch [63/3001], MSE_loss:0.14792 epoch [64/3001], MSE_loss:0.14104 epoch [65/3001], MSE_loss:0.13713 epoch [66/3001], MSE_loss:0.09667 epoch [67/3001], MSE_loss:0.19998 epoch [68/3001], MSE_loss:0.13405 epoch [69/3001], MSE_loss:0.20581 epoch [70/3001], MSE_loss:0.15449 epoch [71/3001], MSE_loss:0.14403 epoch [72/3001], MSE_loss:0.18129 epoch [73/3001], MSE_loss:0.13224 epoch [74/3001], MSE_loss:0.14803 epoch [75/3001], MSE_loss:0.13486 epoch [76/3001], MSE_loss:0.11718 epoch [77/3001], MSE_loss:0.18599 epoch [78/3001], MSE_loss:0.18145 epoch [79/3001], MSE_loss:0.19709 epoch [80/3001], MSE_loss:0.15749 epoch [81/3001], MSE_loss:0.22292 epoch [82/3001], MSE_loss:0.20083 epoch [83/3001], MSE_loss:0.12421 epoch [84/3001], MSE_loss:0.20385 epoch [85/3001], MSE_loss:0.22795 epoch [86/3001], MSE_loss:0.19708 epoch [87/3001], MSE_loss:0.15954 epoch [88/3001], MSE_loss:0.16493 epoch [89/3001], MSE_loss:0.22265 epoch [90/3001], MSE_loss:0.12221 epoch [91/3001], MSE_loss:0.15470 epoch [92/3001], MSE_loss:0.11887 epoch [93/3001], MSE_loss:0.17445 epoch [94/3001], MSE_loss:0.14605 epoch [95/3001], MSE_loss:0.15622 epoch [96/3001], MSE_loss:0.20959 epoch [97/3001], MSE_loss:0.23634 epoch [98/3001], MSE_loss:0.11330 epoch [99/3001], MSE_loss:0.14998 epoch [100/3001], MSE_loss:0.11491 epoch [101/3001], MSE_loss:0.13695 epoch [102/3001], MSE_loss:0.16632 epoch [103/3001], MSE_loss:0.09774 epoch [104/3001], MSE_loss:0.10255 epoch [105/3001], MSE_loss:0.07000 epoch [106/3001], MSE_loss:0.08687 epoch [107/3001], MSE_loss:0.10049 epoch [108/3001], MSE_loss:0.07502 epoch [109/3001], MSE_loss:0.09868 epoch [110/3001], MSE_loss:0.09831 epoch [111/3001], MSE_loss:0.09446 epoch [112/3001], MSE_loss:0.10187 epoch [113/3001], MSE_loss:0.14813 epoch [114/3001], MSE_loss:0.11191 epoch [115/3001], MSE_loss:0.11174 epoch [116/3001], MSE_loss:0.12695 epoch [117/3001], MSE_loss:0.10212 epoch [118/3001], MSE_loss:0.12675 epoch [119/3001], MSE_loss:0.10460 epoch [120/3001], MSE_loss:0.10706 epoch [121/3001], MSE_loss:0.09759 epoch [122/3001], MSE_loss:0.09331 epoch [123/3001], MSE_loss:0.06839 epoch [124/3001], MSE_loss:0.10774 epoch [125/3001], MSE_loss:0.11071 epoch [126/3001], MSE_loss:0.09123 epoch [127/3001], MSE_loss:0.09091 epoch [128/3001], MSE_loss:0.10052 epoch [129/3001], MSE_loss:0.08184 epoch [130/3001], MSE_loss:0.08323 epoch [131/3001], MSE_loss:0.12461 epoch [132/3001], MSE_loss:0.07205 epoch [133/3001], MSE_loss:0.08078 epoch [134/3001], MSE_loss:0.08037 epoch [135/3001], MSE_loss:0.12262 epoch [136/3001], MSE_loss:0.10158 epoch [137/3001], MSE_loss:0.14417 epoch [138/3001], MSE_loss:0.08057 epoch [139/3001], MSE_loss:0.08320 epoch [140/3001], MSE_loss:0.08992 epoch [141/3001], MSE_loss:0.08405 epoch [142/3001], MSE_loss:0.11995 epoch [143/3001], MSE_loss:0.06890 epoch [144/3001], MSE_loss:0.07329 epoch [145/3001], MSE_loss:0.08579 epoch [146/3001], MSE_loss:0.07226 epoch [147/3001], MSE_loss:0.07433 epoch [148/3001], MSE_loss:0.10875 epoch [149/3001], MSE_loss:0.08310 epoch [150/3001], MSE_loss:0.12334 epoch [151/3001], MSE_loss:0.11918 epoch [152/3001], MSE_loss:0.06447 epoch [153/3001], MSE_loss:0.06049 epoch [154/3001], MSE_loss:0.12399 epoch [155/3001], MSE_loss:0.11887 epoch [156/3001], MSE_loss:0.07709 epoch [157/3001], MSE_loss:0.12878 epoch [158/3001], MSE_loss:0.08223 epoch [159/3001], MSE_loss:0.09843 epoch [160/3001], MSE_loss:0.10050 epoch [161/3001], MSE_loss:0.14027 epoch [162/3001], MSE_loss:0.07534 epoch [163/3001], MSE_loss:0.11229 epoch [164/3001], MSE_loss:0.08065 epoch [165/3001], MSE_loss:0.09444 epoch [166/3001], MSE_loss:0.07718 epoch [167/3001], MSE_loss:0.09396 epoch [168/3001], MSE_loss:0.09132 epoch [169/3001], MSE_loss:0.07803 epoch [170/3001], MSE_loss:0.09384 epoch [171/3001], MSE_loss:0.07774 epoch [172/3001], MSE_loss:0.11094 epoch [173/3001], MSE_loss:0.08918 epoch [174/3001], MSE_loss:0.09537 epoch [175/3001], MSE_loss:0.11727 epoch [176/3001], MSE_loss:0.07158 epoch [177/3001], MSE_loss:0.11311 epoch [178/3001], MSE_loss:0.08416 epoch [179/3001], MSE_loss:0.06889 epoch [180/3001], MSE_loss:0.09038 epoch [181/3001], MSE_loss:0.09224 epoch [182/3001], MSE_loss:0.08522 epoch [183/3001], MSE_loss:0.11219 epoch [184/3001], MSE_loss:0.07284 epoch [185/3001], MSE_loss:0.11540 epoch [186/3001], MSE_loss:0.09058 epoch [187/3001], MSE_loss:0.09413 epoch [188/3001], MSE_loss:0.08977 epoch [189/3001], MSE_loss:0.08275 epoch [190/3001], MSE_loss:0.08018 epoch [191/3001], MSE_loss:0.07986 epoch [192/3001], MSE_loss:0.11370 epoch [193/3001], MSE_loss:0.07585 epoch [194/3001], MSE_loss:0.07591 epoch [195/3001], MSE_loss:0.08276 epoch [196/3001], MSE_loss:0.11903 epoch [197/3001], MSE_loss:0.08819 epoch [198/3001], MSE_loss:0.11207 epoch [199/3001], MSE_loss:0.07256 epoch [200/3001], MSE_loss:0.13273 epoch [201/3001], MSE_loss:0.13169 epoch [202/3001], MSE_loss:0.06617 epoch [203/3001], MSE_loss:0.06558 epoch [204/3001], MSE_loss:0.10204 epoch [205/3001], MSE_loss:0.08030 epoch [206/3001], MSE_loss:0.09607 epoch [207/3001], MSE_loss:0.12490 epoch [208/3001], MSE_loss:0.11547 epoch [209/3001], MSE_loss:0.08177 epoch [210/3001], MSE_loss:0.10259 epoch [211/3001], MSE_loss:0.09594 epoch [212/3001], MSE_loss:0.09732 epoch [213/3001], MSE_loss:0.13928 epoch [214/3001], MSE_loss:0.12431 epoch [215/3001], MSE_loss:0.09513 epoch [216/3001], MSE_loss:0.12070 epoch [217/3001], MSE_loss:0.09657 epoch [218/3001], MSE_loss:0.08707 epoch [219/3001], MSE_loss:0.08122 epoch [220/3001], MSE_loss:0.08129 epoch [221/3001], MSE_loss:0.09308 epoch [222/3001], MSE_loss:0.10107 epoch [223/3001], MSE_loss:0.08692 epoch [224/3001], MSE_loss:0.09115 epoch [225/3001], MSE_loss:0.10124 epoch [226/3001], MSE_loss:0.07872 epoch [227/3001], MSE_loss:0.10010 epoch [228/3001], MSE_loss:0.08865 epoch [229/3001], MSE_loss:0.10946 epoch [230/3001], MSE_loss:0.10194 epoch [231/3001], MSE_loss:0.09215 epoch [232/3001], MSE_loss:0.10317 epoch [233/3001], MSE_loss:0.08578 epoch [234/3001], MSE_loss:0.11982 epoch [235/3001], MSE_loss:0.12556 epoch [236/3001], MSE_loss:0.09453 epoch [237/3001], MSE_loss:0.05768 epoch [238/3001], MSE_loss:0.08122 epoch [239/3001], MSE_loss:0.10520 epoch [240/3001], MSE_loss:0.10139 epoch [241/3001], MSE_loss:0.09664 epoch [242/3001], MSE_loss:0.07851 epoch [243/3001], MSE_loss:0.07398 epoch [244/3001], MSE_loss:0.08709 epoch [245/3001], MSE_loss:0.11020 epoch [246/3001], MSE_loss:0.13551 epoch [247/3001], MSE_loss:0.08048 epoch [248/3001], MSE_loss:0.08777 epoch [249/3001], MSE_loss:0.06548 epoch [250/3001], MSE_loss:0.10259 epoch [251/3001], MSE_loss:0.07068 epoch [252/3001], MSE_loss:0.11680 epoch [253/3001], MSE_loss:0.07739 epoch [254/3001], MSE_loss:0.11486 epoch [255/3001], MSE_loss:0.11099 epoch [256/3001], MSE_loss:0.07847 epoch [257/3001], MSE_loss:0.09482 epoch [258/3001], MSE_loss:0.07399 epoch [259/3001], MSE_loss:0.07280 epoch [260/3001], MSE_loss:0.10673 epoch [261/3001], MSE_loss:0.07134 epoch [262/3001], MSE_loss:0.07945 epoch [263/3001], MSE_loss:0.07724 epoch [264/3001], MSE_loss:0.06354 epoch [265/3001], MSE_loss:0.10220 epoch [266/3001], MSE_loss:0.14047 epoch [267/3001], MSE_loss:0.10124 epoch [268/3001], MSE_loss:0.07652 epoch [269/3001], MSE_loss:0.08938 epoch [270/3001], MSE_loss:0.13493 epoch [271/3001], MSE_loss:0.16110 epoch [272/3001], MSE_loss:0.14134 epoch [273/3001], MSE_loss:0.10849 epoch [274/3001], MSE_loss:0.10462 epoch [275/3001], MSE_loss:0.10061 epoch [276/3001], MSE_loss:0.11809 epoch [277/3001], MSE_loss:0.08515 epoch [278/3001], MSE_loss:0.08837 epoch [279/3001], MSE_loss:0.09764 epoch [280/3001], MSE_loss:0.08761 epoch [281/3001], MSE_loss:0.14317 epoch [282/3001], MSE_loss:0.08057 epoch [283/3001], MSE_loss:0.08846 epoch [284/3001], MSE_loss:0.12240 epoch [285/3001], MSE_loss:0.09681 epoch [286/3001], MSE_loss:0.07208 epoch [287/3001], MSE_loss:0.13419 epoch [288/3001], MSE_loss:0.09330 epoch [289/3001], MSE_loss:0.12136 epoch [290/3001], MSE_loss:0.09739 epoch [291/3001], MSE_loss:0.08345 epoch [292/3001], MSE_loss:0.07300 epoch [293/3001], MSE_loss:0.09004 epoch [294/3001], MSE_loss:0.11535 epoch [295/3001], MSE_loss:0.07453 epoch [296/3001], MSE_loss:0.08768 epoch [297/3001], MSE_loss:0.09158 epoch [298/3001], MSE_loss:0.07255 epoch [299/3001], MSE_loss:0.11136 epoch [300/3001], MSE_loss:0.08680 epoch [301/3001], MSE_loss:0.07105 epoch [302/3001], MSE_loss:0.07114 epoch [303/3001], MSE_loss:0.07783 epoch [304/3001], MSE_loss:0.09322 epoch [305/3001], MSE_loss:0.09985 epoch [306/3001], MSE_loss:0.06043 epoch [307/3001], MSE_loss:0.07831 epoch [308/3001], MSE_loss:0.10886 epoch [309/3001], MSE_loss:0.08560 epoch [310/3001], MSE_loss:0.07420 epoch [311/3001], MSE_loss:0.09306 epoch [312/3001], MSE_loss:0.08301 epoch [313/3001], MSE_loss:0.08226 epoch [314/3001], MSE_loss:0.08022 epoch [315/3001], MSE_loss:0.08951 epoch [316/3001], MSE_loss:0.07957 epoch [317/3001], MSE_loss:0.14401 epoch [318/3001], MSE_loss:0.11317 epoch [319/3001], MSE_loss:0.11539 epoch [320/3001], MSE_loss:0.10858 epoch [321/3001], MSE_loss:0.09001 epoch [322/3001], MSE_loss:0.10366 epoch [323/3001], MSE_loss:0.14750 epoch [324/3001], MSE_loss:0.12281 epoch [325/3001], MSE_loss:0.10550 epoch [326/3001], MSE_loss:0.09778 epoch [327/3001], MSE_loss:0.05147 epoch [328/3001], MSE_loss:0.09885 epoch [329/3001], MSE_loss:0.07897 epoch [330/3001], MSE_loss:0.10653 epoch [331/3001], MSE_loss:0.08275 epoch [332/3001], MSE_loss:0.13080 epoch [333/3001], MSE_loss:0.11866 epoch [334/3001], MSE_loss:0.15916 epoch [335/3001], MSE_loss:0.10282 epoch [336/3001], MSE_loss:0.08784 epoch [337/3001], MSE_loss:0.08537 epoch [338/3001], MSE_loss:0.10756 epoch [339/3001], MSE_loss:0.09111 epoch [340/3001], MSE_loss:0.09670 epoch [341/3001], MSE_loss:0.09720 epoch [342/3001], MSE_loss:0.15869 epoch [343/3001], MSE_loss:0.11982 epoch [344/3001], MSE_loss:0.11125 epoch [345/3001], MSE_loss:0.08778 epoch [346/3001], MSE_loss:0.13624 epoch [347/3001], MSE_loss:0.09656 epoch [348/3001], MSE_loss:0.08473 epoch [349/3001], MSE_loss:0.09558 epoch [350/3001], MSE_loss:0.11489 epoch [351/3001], MSE_loss:0.09191 epoch [352/3001], MSE_loss:0.10606 epoch [353/3001], MSE_loss:0.09260 epoch [354/3001], MSE_loss:0.08160 epoch [355/3001], MSE_loss:0.09236 epoch [356/3001], MSE_loss:0.08874 epoch [357/3001], MSE_loss:0.07336 epoch [358/3001], MSE_loss:0.09076 epoch [359/3001], MSE_loss:0.09882 epoch [360/3001], MSE_loss:0.06466 epoch [361/3001], MSE_loss:0.12065 epoch [362/3001], MSE_loss:0.08733 epoch [363/3001], MSE_loss:0.08449 epoch [364/3001], MSE_loss:0.09269 epoch [365/3001], MSE_loss:0.11927 epoch [366/3001], MSE_loss:0.09492 epoch [367/3001], MSE_loss:0.09637 epoch [368/3001], MSE_loss:0.10056 epoch [369/3001], MSE_loss:0.07901 epoch [370/3001], MSE_loss:0.07318 epoch [371/3001], MSE_loss:0.11932 epoch [372/3001], MSE_loss:0.06606 epoch [373/3001], MSE_loss:0.08038 epoch [374/3001], MSE_loss:0.09910 epoch [375/3001], MSE_loss:0.13777 epoch [376/3001], MSE_loss:0.06823 epoch [377/3001], MSE_loss:0.06353 epoch [378/3001], MSE_loss:0.11696 epoch [379/3001], MSE_loss:0.09992 epoch [380/3001], MSE_loss:0.11806 epoch [381/3001], MSE_loss:0.12937 epoch [382/3001], MSE_loss:0.08123 epoch [383/3001], MSE_loss:0.09759 epoch [384/3001], MSE_loss:0.11573 epoch [385/3001], MSE_loss:0.08369 epoch [386/3001], MSE_loss:0.07412 epoch [387/3001], MSE_loss:0.12753 epoch [388/3001], MSE_loss:0.07563 epoch [389/3001], MSE_loss:0.06167 epoch [390/3001], MSE_loss:0.11667 epoch [391/3001], MSE_loss:0.08568 epoch [392/3001], MSE_loss:0.08144 epoch [393/3001], MSE_loss:0.08630 epoch [394/3001], MSE_loss:0.10692 epoch [395/3001], MSE_loss:0.07260 epoch [396/3001], MSE_loss:0.08368 epoch [397/3001], MSE_loss:0.09562 epoch [398/3001], MSE_loss:0.09326 epoch [399/3001], MSE_loss:0.06276 epoch [400/3001], MSE_loss:0.11824 epoch [401/3001], MSE_loss:0.08174 epoch [402/3001], MSE_loss:0.05971 epoch [403/3001], MSE_loss:0.12007 epoch [404/3001], MSE_loss:0.06501 epoch [405/3001], MSE_loss:0.08720 epoch [406/3001], MSE_loss:0.08032 epoch [407/3001], MSE_loss:0.08213 epoch [408/3001], MSE_loss:0.10264 epoch [409/3001], MSE_loss:0.12068 epoch [410/3001], MSE_loss:0.10111 epoch [411/3001], MSE_loss:0.06520 epoch [412/3001], MSE_loss:0.11403 epoch [413/3001], MSE_loss:0.10413 epoch [414/3001], MSE_loss:0.08800 epoch [415/3001], MSE_loss:0.16073 epoch [416/3001], MSE_loss:0.09356 epoch [417/3001], MSE_loss:0.11735 epoch [418/3001], MSE_loss:0.10786 epoch [419/3001], MSE_loss:0.09879 epoch [420/3001], MSE_loss:0.12068 epoch [421/3001], MSE_loss:0.11120 epoch [422/3001], MSE_loss:0.08133 epoch [423/3001], MSE_loss:0.09908 epoch [424/3001], MSE_loss:0.09453 epoch [425/3001], MSE_loss:0.08606 epoch [426/3001], MSE_loss:0.09250 epoch [427/3001], MSE_loss:0.07923 epoch [428/3001], MSE_loss:0.07905 epoch [429/3001], MSE_loss:0.09559 epoch [430/3001], MSE_loss:0.08314 epoch [431/3001], MSE_loss:0.08861 epoch [432/3001], MSE_loss:0.09409 epoch [433/3001], MSE_loss:0.09314 epoch [434/3001], MSE_loss:0.12926 epoch [435/3001], MSE_loss:0.09328 epoch [436/3001], MSE_loss:0.07615 epoch [437/3001], MSE_loss:0.12015 epoch [438/3001], MSE_loss:0.08026 epoch [439/3001], MSE_loss:0.10877 epoch [440/3001], MSE_loss:0.09202 epoch [441/3001], MSE_loss:0.12107 epoch [442/3001], MSE_loss:0.10533 epoch [443/3001], MSE_loss:0.08722 epoch [444/3001], MSE_loss:0.08609 epoch [445/3001], MSE_loss:0.10963 epoch [446/3001], MSE_loss:0.05449 epoch [447/3001], MSE_loss:0.08026 epoch [448/3001], MSE_loss:0.09687 epoch [449/3001], MSE_loss:0.08512 epoch [450/3001], MSE_loss:0.06704 epoch [451/3001], MSE_loss:0.07671 epoch [452/3001], MSE_loss:0.09809 epoch [453/3001], MSE_loss:0.09114 epoch [454/3001], MSE_loss:0.10353 epoch [455/3001], MSE_loss:0.13080 epoch [456/3001], MSE_loss:0.06165 epoch [457/3001], MSE_loss:0.12407 epoch [458/3001], MSE_loss:0.11566 epoch [459/3001], MSE_loss:0.09326 epoch [460/3001], MSE_loss:0.07549 epoch [461/3001], MSE_loss:0.09661 epoch [462/3001], MSE_loss:0.06855 epoch [463/3001], MSE_loss:0.07743 epoch [464/3001], MSE_loss:0.09848 epoch [465/3001], MSE_loss:0.10753 epoch [466/3001], MSE_loss:0.07787 epoch [467/3001], MSE_loss:0.11211 epoch [468/3001], MSE_loss:0.11590 epoch [469/3001], MSE_loss:0.10143 epoch [470/3001], MSE_loss:0.08991 epoch [471/3001], MSE_loss:0.09011 epoch [472/3001], MSE_loss:0.07048 epoch [473/3001], MSE_loss:0.09384 epoch [474/3001], MSE_loss:0.06511 epoch [475/3001], MSE_loss:0.10348 epoch [476/3001], MSE_loss:0.08773 epoch [477/3001], MSE_loss:0.08186 epoch [478/3001], MSE_loss:0.05922 epoch [479/3001], MSE_loss:0.08061 epoch [480/3001], MSE_loss:0.08311 epoch [481/3001], MSE_loss:0.09648 epoch [482/3001], MSE_loss:0.11609 epoch [483/3001], MSE_loss:0.08534 epoch [484/3001], MSE_loss:0.05743 epoch [485/3001], MSE_loss:0.08090 epoch [486/3001], MSE_loss:0.08988 epoch [487/3001], MSE_loss:0.12743 epoch [488/3001], MSE_loss:0.09772 epoch [489/3001], MSE_loss:0.10495 epoch [490/3001], MSE_loss:0.16594 epoch [491/3001], MSE_loss:0.10883 epoch [492/3001], MSE_loss:0.09088 epoch [493/3001], MSE_loss:0.06740 epoch [494/3001], MSE_loss:0.07047 epoch [495/3001], MSE_loss:0.08840 epoch [496/3001], MSE_loss:0.09252 epoch [497/3001], MSE_loss:0.06516 epoch [498/3001], MSE_loss:0.08384 epoch [499/3001], MSE_loss:0.11041 epoch [500/3001], MSE_loss:0.08494 epoch [501/3001], MSE_loss:0.07953 epoch [502/3001], MSE_loss:0.11146 epoch [503/3001], MSE_loss:0.09434 epoch [504/3001], MSE_loss:0.07191 epoch [505/3001], MSE_loss:0.07419 epoch [506/3001], MSE_loss:0.07552 epoch [507/3001], MSE_loss:0.10301 epoch [508/3001], MSE_loss:0.08656 epoch [509/3001], MSE_loss:0.07916 epoch [510/3001], MSE_loss:0.08684 epoch [511/3001], MSE_loss:0.09355 epoch [512/3001], MSE_loss:0.10454 epoch [513/3001], MSE_loss:0.08446 epoch [514/3001], MSE_loss:0.07815 epoch [515/3001], MSE_loss:0.08669 epoch [516/3001], MSE_loss:0.10610 epoch [517/3001], MSE_loss:0.09967 epoch [518/3001], MSE_loss:0.06759 epoch [519/3001], MSE_loss:0.07711 epoch [520/3001], MSE_loss:0.08121 epoch [521/3001], MSE_loss:0.09073 epoch [522/3001], MSE_loss:0.11427 epoch [523/3001], MSE_loss:0.09820 epoch [524/3001], MSE_loss:0.08866 epoch [525/3001], MSE_loss:0.07843 epoch [526/3001], MSE_loss:0.08182 epoch [527/3001], MSE_loss:0.11459 epoch [528/3001], MSE_loss:0.12956 epoch [529/3001], MSE_loss:0.09714 epoch [530/3001], MSE_loss:0.12269 epoch [531/3001], MSE_loss:0.10613 epoch [532/3001], MSE_loss:0.10036 epoch [533/3001], MSE_loss:0.09844 epoch [534/3001], MSE_loss:0.08295 epoch [535/3001], MSE_loss:0.11736 epoch [536/3001], MSE_loss:0.14586 epoch [537/3001], MSE_loss:0.10493 epoch [538/3001], MSE_loss:0.13754 epoch [539/3001], MSE_loss:0.09411 epoch [540/3001], MSE_loss:0.10252 epoch [541/3001], MSE_loss:0.11208 epoch [542/3001], MSE_loss:0.10011 epoch [543/3001], MSE_loss:0.09162 epoch [544/3001], MSE_loss:0.07679 epoch [545/3001], MSE_loss:0.12039 epoch [546/3001], MSE_loss:0.10349 epoch [547/3001], MSE_loss:0.09623 epoch [548/3001], MSE_loss:0.07169 epoch [549/3001], MSE_loss:0.17445 epoch [550/3001], MSE_loss:0.09749 epoch [551/3001], MSE_loss:0.10028 epoch [552/3001], MSE_loss:0.08884 epoch [553/3001], MSE_loss:0.12381 epoch [554/3001], MSE_loss:0.09448 epoch [555/3001], MSE_loss:0.10085 epoch [556/3001], MSE_loss:0.08368 epoch [557/3001], MSE_loss:0.07702 epoch [558/3001], MSE_loss:0.06221 epoch [559/3001], MSE_loss:0.12469 epoch [560/3001], MSE_loss:0.08494 epoch [561/3001], MSE_loss:0.09580 epoch [562/3001], MSE_loss:0.09976 epoch [563/3001], MSE_loss:0.05917 epoch [564/3001], MSE_loss:0.08952 epoch [565/3001], MSE_loss:0.13424 epoch [566/3001], MSE_loss:0.11499 epoch [567/3001], MSE_loss:0.06803 epoch [568/3001], MSE_loss:0.08458 epoch [569/3001], MSE_loss:0.08387 epoch [570/3001], MSE_loss:0.10883 epoch [571/3001], MSE_loss:0.08013 epoch [572/3001], MSE_loss:0.07846 epoch [573/3001], MSE_loss:0.07685 epoch [574/3001], MSE_loss:0.09529 epoch [575/3001], MSE_loss:0.11668 epoch [576/3001], MSE_loss:0.08105 epoch [577/3001], MSE_loss:0.09221 epoch [578/3001], MSE_loss:0.09281 epoch [579/3001], MSE_loss:0.12946 epoch [580/3001], MSE_loss:0.10342 epoch [581/3001], MSE_loss:0.08588 epoch [582/3001], MSE_loss:0.06477 epoch [583/3001], MSE_loss:0.07796 epoch [584/3001], MSE_loss:0.06914 epoch [585/3001], MSE_loss:0.09581 epoch [586/3001], MSE_loss:0.10384 epoch [587/3001], MSE_loss:0.10288 epoch [588/3001], MSE_loss:0.09486 epoch [589/3001], MSE_loss:0.08647 epoch [590/3001], MSE_loss:0.08600 epoch [591/3001], MSE_loss:0.09312 epoch [592/3001], MSE_loss:0.08229 epoch [593/3001], MSE_loss:0.08118 epoch [594/3001], MSE_loss:0.10179 epoch [595/3001], MSE_loss:0.07548 epoch [596/3001], MSE_loss:0.09799 epoch [597/3001], MSE_loss:0.09665 epoch [598/3001], MSE_loss:0.11057 epoch [599/3001], MSE_loss:0.08086 epoch [600/3001], MSE_loss:0.11204 epoch [601/3001], MSE_loss:0.08134 epoch [602/3001], MSE_loss:0.05904 epoch [603/3001], MSE_loss:0.08329 epoch [604/3001], MSE_loss:0.08674 epoch [605/3001], MSE_loss:0.07003 epoch [606/3001], MSE_loss:0.10985 epoch [607/3001], MSE_loss:0.13518 epoch [608/3001], MSE_loss:0.11917 epoch [609/3001], MSE_loss:0.05394 epoch [610/3001], MSE_loss:0.08942 epoch [611/3001], MSE_loss:0.11723 epoch [612/3001], MSE_loss:0.06834 epoch [613/3001], MSE_loss:0.11395 epoch [614/3001], MSE_loss:0.09546 epoch [615/3001], MSE_loss:0.09028 epoch [616/3001], MSE_loss:0.09267 epoch [617/3001], MSE_loss:0.07610 epoch [618/3001], MSE_loss:0.05500 epoch [619/3001], MSE_loss:0.12798 epoch [620/3001], MSE_loss:0.05038 epoch [621/3001], MSE_loss:0.08747 epoch [622/3001], MSE_loss:0.10856 epoch [623/3001], MSE_loss:0.09282 epoch [624/3001], MSE_loss:0.09419 epoch [625/3001], MSE_loss:0.10116 epoch [626/3001], MSE_loss:0.09214 epoch [627/3001], MSE_loss:0.07877 epoch [628/3001], MSE_loss:0.08820 epoch [629/3001], MSE_loss:0.07884 epoch [630/3001], MSE_loss:0.10052 epoch [631/3001], MSE_loss:0.07201 epoch [632/3001], MSE_loss:0.06817 epoch [633/3001], MSE_loss:0.07644 epoch [634/3001], MSE_loss:0.13983 epoch [635/3001], MSE_loss:0.08108 epoch [636/3001], MSE_loss:0.08450 epoch [637/3001], MSE_loss:0.09881 epoch [638/3001], MSE_loss:0.08610 epoch [639/3001], MSE_loss:0.12756 epoch [640/3001], MSE_loss:0.11738 epoch [641/3001], MSE_loss:0.10767 epoch [642/3001], MSE_loss:0.11178 epoch [643/3001], MSE_loss:0.08197 epoch [644/3001], MSE_loss:0.09949 epoch [645/3001], MSE_loss:0.08514 epoch [646/3001], MSE_loss:0.08242 epoch [647/3001], MSE_loss:0.09901 epoch [648/3001], MSE_loss:0.11687 epoch [649/3001], MSE_loss:0.06334 epoch [650/3001], MSE_loss:0.09062 epoch [651/3001], MSE_loss:0.08782 epoch [652/3001], MSE_loss:0.06768 epoch [653/3001], MSE_loss:0.11278 epoch [654/3001], MSE_loss:0.08665 epoch [655/3001], MSE_loss:0.10053 epoch [656/3001], MSE_loss:0.10557 epoch [657/3001], MSE_loss:0.10200 epoch [658/3001], MSE_loss:0.07011 epoch [659/3001], MSE_loss:0.13429 epoch [660/3001], MSE_loss:0.11076 epoch [661/3001], MSE_loss:0.14213 epoch [662/3001], MSE_loss:0.15194 epoch [663/3001], MSE_loss:0.08273 epoch [664/3001], MSE_loss:0.06894 epoch [665/3001], MSE_loss:0.08239 epoch [666/3001], MSE_loss:0.11316 epoch [667/3001], MSE_loss:0.08405 epoch [668/3001], MSE_loss:0.10570 epoch [669/3001], MSE_loss:0.08623 epoch [670/3001], MSE_loss:0.10373 epoch [671/3001], MSE_loss:0.12051 epoch [672/3001], MSE_loss:0.08385 epoch [673/3001], MSE_loss:0.08456 epoch [674/3001], MSE_loss:0.06939 epoch [675/3001], MSE_loss:0.09437 epoch [676/3001], MSE_loss:0.08013 epoch [677/3001], MSE_loss:0.12054 epoch [678/3001], MSE_loss:0.08578 epoch [679/3001], MSE_loss:0.08124 epoch [680/3001], MSE_loss:0.12198 epoch [681/3001], MSE_loss:0.07549 epoch [682/3001], MSE_loss:0.09055 epoch [683/3001], MSE_loss:0.11222 epoch [684/3001], MSE_loss:0.11013 epoch [685/3001], MSE_loss:0.06958 epoch [686/3001], MSE_loss:0.09924 epoch [687/3001], MSE_loss:0.12344 epoch [688/3001], MSE_loss:0.09705 epoch [689/3001], MSE_loss:0.12521 epoch [690/3001], MSE_loss:0.07411 epoch [691/3001], MSE_loss:0.05185 epoch [692/3001], MSE_loss:0.10306 epoch [693/3001], MSE_loss:0.09693 epoch [694/3001], MSE_loss:0.09333 epoch [695/3001], MSE_loss:0.09970 epoch [696/3001], MSE_loss:0.08059 epoch [697/3001], MSE_loss:0.12793 epoch [698/3001], MSE_loss:0.10237 epoch [699/3001], MSE_loss:0.10402 epoch [700/3001], MSE_loss:0.10164 epoch [701/3001], MSE_loss:0.09266 epoch [702/3001], MSE_loss:0.09669 epoch [703/3001], MSE_loss:0.09658 epoch [704/3001], MSE_loss:0.10742 epoch [705/3001], MSE_loss:0.06884 epoch [706/3001], MSE_loss:0.12633 epoch [707/3001], MSE_loss:0.10455 epoch [708/3001], MSE_loss:0.07864 epoch [709/3001], MSE_loss:0.12162 epoch [710/3001], MSE_loss:0.13198 epoch [711/3001], MSE_loss:0.06198 epoch [712/3001], MSE_loss:0.09118 epoch [713/3001], MSE_loss:0.10346 epoch [714/3001], MSE_loss:0.06761 epoch [715/3001], MSE_loss:0.07813 epoch [716/3001], MSE_loss:0.07083 epoch [717/3001], MSE_loss:0.10104 epoch [718/3001], MSE_loss:0.09166 epoch [719/3001], MSE_loss:0.10724 epoch [720/3001], MSE_loss:0.06955 epoch [721/3001], MSE_loss:0.08777 epoch [722/3001], MSE_loss:0.08697 epoch [723/3001], MSE_loss:0.12550 epoch [724/3001], MSE_loss:0.09043 epoch [725/3001], MSE_loss:0.11051 epoch [726/3001], MSE_loss:0.06284 epoch [727/3001], MSE_loss:0.09472 epoch [728/3001], MSE_loss:0.06529 epoch [729/3001], MSE_loss:0.08955 epoch [730/3001], MSE_loss:0.09525 epoch [731/3001], MSE_loss:0.09736 epoch [732/3001], MSE_loss:0.11921 epoch [733/3001], MSE_loss:0.07161 epoch [734/3001], MSE_loss:0.10059 epoch [735/3001], MSE_loss:0.09784 epoch [736/3001], MSE_loss:0.07008 epoch [737/3001], MSE_loss:0.09773 epoch [738/3001], MSE_loss:0.10169 epoch [739/3001], MSE_loss:0.08266 epoch [740/3001], MSE_loss:0.08097 epoch [741/3001], MSE_loss:0.07545 epoch [742/3001], MSE_loss:0.06728 epoch [743/3001], MSE_loss:0.11122 epoch [744/3001], MSE_loss:0.10361 epoch [745/3001], MSE_loss:0.06434 epoch [746/3001], MSE_loss:0.07109 epoch [747/3001], MSE_loss:0.10234 epoch [748/3001], MSE_loss:0.10238 epoch [749/3001], MSE_loss:0.08828 epoch [750/3001], MSE_loss:0.09344 epoch [751/3001], MSE_loss:0.08916 epoch [752/3001], MSE_loss:0.10661 epoch [753/3001], MSE_loss:0.12409 epoch [754/3001], MSE_loss:0.08941 epoch [755/3001], MSE_loss:0.10554 epoch [756/3001], MSE_loss:0.06815 epoch [757/3001], MSE_loss:0.08217 epoch [758/3001], MSE_loss:0.11441 epoch [759/3001], MSE_loss:0.11978 epoch [760/3001], MSE_loss:0.07594 epoch [761/3001], MSE_loss:0.12972 epoch [762/3001], MSE_loss:0.06205 epoch [763/3001], MSE_loss:0.07639 epoch [764/3001], MSE_loss:0.06825 epoch [765/3001], MSE_loss:0.07626 epoch [766/3001], MSE_loss:0.09506 epoch [767/3001], MSE_loss:0.09901 epoch [768/3001], MSE_loss:0.07817 epoch [769/3001], MSE_loss:0.12622 epoch [770/3001], MSE_loss:0.08159 epoch [771/3001], MSE_loss:0.13972 epoch [772/3001], MSE_loss:0.09742 epoch [773/3001], MSE_loss:0.06919 epoch [774/3001], MSE_loss:0.11119 epoch [775/3001], MSE_loss:0.09487 epoch [776/3001], MSE_loss:0.09674 epoch [777/3001], MSE_loss:0.08832 epoch [778/3001], MSE_loss:0.09652 epoch [779/3001], MSE_loss:0.09754 epoch [780/3001], MSE_loss:0.09515 epoch [781/3001], MSE_loss:0.12841 epoch [782/3001], MSE_loss:0.13725 epoch [783/3001], MSE_loss:0.10420 epoch [784/3001], MSE_loss:0.11084 epoch [785/3001], MSE_loss:0.10013 epoch [786/3001], MSE_loss:0.11789 epoch [787/3001], MSE_loss:0.10533 epoch [788/3001], MSE_loss:0.08693 epoch [789/3001], MSE_loss:0.10248 epoch [790/3001], MSE_loss:0.05949 epoch [791/3001], MSE_loss:0.08191 epoch [792/3001], MSE_loss:0.12638 epoch [793/3001], MSE_loss:0.08732 epoch [794/3001], MSE_loss:0.09283 epoch [795/3001], MSE_loss:0.11697 epoch [796/3001], MSE_loss:0.06898 epoch [797/3001], MSE_loss:0.08972 epoch [798/3001], MSE_loss:0.14355 epoch [799/3001], MSE_loss:0.08586 epoch [800/3001], MSE_loss:0.10204 epoch [801/3001], MSE_loss:0.10766 epoch [802/3001], MSE_loss:0.11308 epoch [803/3001], MSE_loss:0.10734 epoch [804/3001], MSE_loss:0.07665 epoch [805/3001], MSE_loss:0.07957 epoch [806/3001], MSE_loss:0.13142 epoch [807/3001], MSE_loss:0.11903 epoch [808/3001], MSE_loss:0.10646 epoch [809/3001], MSE_loss:0.07198 epoch [810/3001], MSE_loss:0.10047 epoch [811/3001], MSE_loss:0.07166 epoch [812/3001], MSE_loss:0.10721 epoch [813/3001], MSE_loss:0.09922 epoch [814/3001], MSE_loss:0.06787 epoch [815/3001], MSE_loss:0.11925 epoch [816/3001], MSE_loss:0.07683 epoch [817/3001], MSE_loss:0.13985 epoch [818/3001], MSE_loss:0.11574 epoch [819/3001], MSE_loss:0.08667 epoch [820/3001], MSE_loss:0.07276 epoch [821/3001], MSE_loss:0.10584 epoch [822/3001], MSE_loss:0.15478 epoch [823/3001], MSE_loss:0.09845 epoch [824/3001], MSE_loss:0.11517 epoch [825/3001], MSE_loss:0.09726 epoch [826/3001], MSE_loss:0.13061 epoch [827/3001], MSE_loss:0.08433 epoch [828/3001], MSE_loss:0.07893 epoch [829/3001], MSE_loss:0.06774 epoch [830/3001], MSE_loss:0.08211 epoch [831/3001], MSE_loss:0.10258 epoch [832/3001], MSE_loss:0.08068 epoch [833/3001], MSE_loss:0.07399 epoch [834/3001], MSE_loss:0.08363 epoch [835/3001], MSE_loss:0.09916 epoch [836/3001], MSE_loss:0.11723 epoch [837/3001], MSE_loss:0.09954 epoch [838/3001], MSE_loss:0.11002 epoch [839/3001], MSE_loss:0.05905 epoch [840/3001], MSE_loss:0.06812 epoch [841/3001], MSE_loss:0.09062 epoch [842/3001], MSE_loss:0.07170 epoch [843/3001], MSE_loss:0.07682 epoch [844/3001], MSE_loss:0.10657 epoch [845/3001], MSE_loss:0.06119 epoch [846/3001], MSE_loss:0.06796 epoch [847/3001], MSE_loss:0.09812 epoch [848/3001], MSE_loss:0.11222 epoch [849/3001], MSE_loss:0.06274 epoch [850/3001], MSE_loss:0.08660 epoch [851/3001], MSE_loss:0.06470 epoch [852/3001], MSE_loss:0.09736 epoch [853/3001], MSE_loss:0.07525 epoch [854/3001], MSE_loss:0.10657 epoch [855/3001], MSE_loss:0.11412 epoch [856/3001], MSE_loss:0.10361 epoch [857/3001], MSE_loss:0.08492 epoch [858/3001], MSE_loss:0.07696 epoch [859/3001], MSE_loss:0.10711 epoch [860/3001], MSE_loss:0.09082 epoch [861/3001], MSE_loss:0.09969 epoch [862/3001], MSE_loss:0.07992 epoch [863/3001], MSE_loss:0.08674 epoch [864/3001], MSE_loss:0.09638 epoch [865/3001], MSE_loss:0.06565 epoch [866/3001], MSE_loss:0.10884 epoch [867/3001], MSE_loss:0.11586 epoch [868/3001], MSE_loss:0.07845 epoch [869/3001], MSE_loss:0.05188 epoch [870/3001], MSE_loss:0.08513 epoch [871/3001], MSE_loss:0.14258 epoch [872/3001], MSE_loss:0.10429 epoch [873/3001], MSE_loss:0.12488 epoch [874/3001], MSE_loss:0.09089 epoch [875/3001], MSE_loss:0.07727 epoch [876/3001], MSE_loss:0.09890 epoch [877/3001], MSE_loss:0.07859 epoch [878/3001], MSE_loss:0.12814 epoch [879/3001], MSE_loss:0.08950 epoch [880/3001], MSE_loss:0.08783 epoch [881/3001], MSE_loss:0.08485 epoch [882/3001], MSE_loss:0.07126 epoch [883/3001], MSE_loss:0.10250 epoch [884/3001], MSE_loss:0.08035 epoch [885/3001], MSE_loss:0.09968 epoch [886/3001], MSE_loss:0.09365 epoch [887/3001], MSE_loss:0.08252 epoch [888/3001], MSE_loss:0.10142 epoch [889/3001], MSE_loss:0.07892 epoch [890/3001], MSE_loss:0.07717 epoch [891/3001], MSE_loss:0.07830 epoch [892/3001], MSE_loss:0.07886 epoch [893/3001], MSE_loss:0.12277 epoch [894/3001], MSE_loss:0.08535 epoch [895/3001], MSE_loss:0.07865 epoch [896/3001], MSE_loss:0.08068 epoch [897/3001], MSE_loss:0.09909 epoch [898/3001], MSE_loss:0.13721 epoch [899/3001], MSE_loss:0.09115 epoch [900/3001], MSE_loss:0.14170 epoch [901/3001], MSE_loss:0.09969 epoch [902/3001], MSE_loss:0.08917 epoch [903/3001], MSE_loss:0.08161 epoch [904/3001], MSE_loss:0.08867 epoch [905/3001], MSE_loss:0.09607 epoch [906/3001], MSE_loss:0.06461 epoch [907/3001], MSE_loss:0.08694 epoch [908/3001], MSE_loss:0.13081 epoch [909/3001], MSE_loss:0.09607 epoch [910/3001], MSE_loss:0.11204 epoch [911/3001], MSE_loss:0.09666 epoch [912/3001], MSE_loss:0.10657 epoch [913/3001], MSE_loss:0.09779 epoch [914/3001], MSE_loss:0.10132 epoch [915/3001], MSE_loss:0.05087 epoch [916/3001], MSE_loss:0.08440 epoch [917/3001], MSE_loss:0.05795 epoch [918/3001], MSE_loss:0.14097 epoch [919/3001], MSE_loss:0.08119 epoch [920/3001], MSE_loss:0.08716 epoch [921/3001], MSE_loss:0.09211 epoch [922/3001], MSE_loss:0.06398 epoch [923/3001], MSE_loss:0.09681 epoch [924/3001], MSE_loss:0.13318 epoch [925/3001], MSE_loss:0.09897 epoch [926/3001], MSE_loss:0.07115 epoch [927/3001], MSE_loss:0.13510 epoch [928/3001], MSE_loss:0.09147 epoch [929/3001], MSE_loss:0.07851 epoch [930/3001], MSE_loss:0.07245 epoch [931/3001], MSE_loss:0.05236 epoch [932/3001], MSE_loss:0.05426 epoch [933/3001], MSE_loss:0.09358 epoch [934/3001], MSE_loss:0.16558 epoch [935/3001], MSE_loss:0.10031 epoch [936/3001], MSE_loss:0.06908 epoch [937/3001], MSE_loss:0.09238 epoch [938/3001], MSE_loss:0.13120 epoch [939/3001], MSE_loss:0.11988 epoch [940/3001], MSE_loss:0.13506 epoch [941/3001], MSE_loss:0.08952 epoch [942/3001], MSE_loss:0.09533 epoch [943/3001], MSE_loss:0.15597 epoch [944/3001], MSE_loss:0.06806 epoch [945/3001], MSE_loss:0.10645 epoch [946/3001], MSE_loss:0.11690 epoch [947/3001], MSE_loss:0.10796 epoch [948/3001], MSE_loss:0.08061 epoch [949/3001], MSE_loss:0.09274 epoch [950/3001], MSE_loss:0.11714 epoch [951/3001], MSE_loss:0.10386 epoch [952/3001], MSE_loss:0.07593 epoch [953/3001], MSE_loss:0.16021 epoch [954/3001], MSE_loss:0.13963 epoch [955/3001], MSE_loss:0.07354 epoch [956/3001], MSE_loss:0.10311 epoch [957/3001], MSE_loss:0.07889 epoch [958/3001], MSE_loss:0.08026 epoch [959/3001], MSE_loss:0.08243 epoch [960/3001], MSE_loss:0.09267 epoch [961/3001], MSE_loss:0.10503 epoch [962/3001], MSE_loss:0.12577 epoch [963/3001], MSE_loss:0.10555 epoch [964/3001], MSE_loss:0.10597 epoch [965/3001], MSE_loss:0.08222 epoch [966/3001], MSE_loss:0.13906 epoch [967/3001], MSE_loss:0.09990 epoch [968/3001], MSE_loss:0.08209 epoch [969/3001], MSE_loss:0.10323 epoch [970/3001], MSE_loss:0.08935 epoch [971/3001], MSE_loss:0.12540 epoch [972/3001], MSE_loss:0.08986 epoch [973/3001], MSE_loss:0.08338 epoch [974/3001], MSE_loss:0.10494 epoch [975/3001], MSE_loss:0.08613 epoch [976/3001], MSE_loss:0.09503 epoch [977/3001], MSE_loss:0.07953 epoch [978/3001], MSE_loss:0.12753 epoch [979/3001], MSE_loss:0.13672 epoch [980/3001], MSE_loss:0.11413 epoch [981/3001], MSE_loss:0.14226 epoch [982/3001], MSE_loss:0.09781 epoch [983/3001], MSE_loss:0.09605 epoch [984/3001], MSE_loss:0.06870 epoch [985/3001], MSE_loss:0.07999 epoch [986/3001], MSE_loss:0.07986 epoch [987/3001], MSE_loss:0.10267 epoch [988/3001], MSE_loss:0.09252 epoch [989/3001], MSE_loss:0.07308 epoch [990/3001], MSE_loss:0.11050 epoch [991/3001], MSE_loss:0.08736 epoch [992/3001], MSE_loss:0.07580 epoch [993/3001], MSE_loss:0.07756 epoch [994/3001], MSE_loss:0.10338 epoch [995/3001], MSE_loss:0.09438 epoch [996/3001], MSE_loss:0.11324 epoch [997/3001], MSE_loss:0.12360 epoch [998/3001], MSE_loss:0.08198 epoch [999/3001], MSE_loss:0.07696 epoch [1000/3001], MSE_loss:0.10087 epoch [1001/3001], MSE_loss:0.09321 epoch [1002/3001], MSE_loss:0.11779 epoch [1003/3001], MSE_loss:0.09746 epoch [1004/3001], MSE_loss:0.12242 epoch [1005/3001], MSE_loss:0.10657 epoch [1006/3001], MSE_loss:0.09382 epoch [1007/3001], MSE_loss:0.07498 epoch [1008/3001], MSE_loss:0.07913 epoch [1009/3001], MSE_loss:0.07931 epoch [1010/3001], MSE_loss:0.09101 epoch [1011/3001], MSE_loss:0.09382 epoch [1012/3001], MSE_loss:0.11795 epoch [1013/3001], MSE_loss:0.10237 epoch [1014/3001], MSE_loss:0.13409 epoch [1015/3001], MSE_loss:0.10025 epoch [1016/3001], MSE_loss:0.06821 epoch [1017/3001], MSE_loss:0.07350 epoch [1018/3001], MSE_loss:0.12416 epoch [1019/3001], MSE_loss:0.15139 epoch [1020/3001], MSE_loss:0.06667 epoch [1021/3001], MSE_loss:0.12136 epoch [1022/3001], MSE_loss:0.11039 epoch [1023/3001], MSE_loss:0.08716 epoch [1024/3001], MSE_loss:0.08751 epoch [1025/3001], MSE_loss:0.09126 epoch [1026/3001], MSE_loss:0.09662 epoch [1027/3001], MSE_loss:0.09070 epoch [1028/3001], MSE_loss:0.09792 epoch [1029/3001], MSE_loss:0.08426 epoch [1030/3001], MSE_loss:0.10286 epoch [1031/3001], MSE_loss:0.06381 epoch [1032/3001], MSE_loss:0.10708 epoch [1033/3001], MSE_loss:0.07106 epoch [1034/3001], MSE_loss:0.09301 epoch [1035/3001], MSE_loss:0.09623 epoch [1036/3001], MSE_loss:0.12984 epoch [1037/3001], MSE_loss:0.06631 epoch [1038/3001], MSE_loss:0.06173 epoch [1039/3001], MSE_loss:0.10524 epoch [1040/3001], MSE_loss:0.08221 epoch [1041/3001], MSE_loss:0.11025 epoch [1042/3001], MSE_loss:0.09327 epoch [1043/3001], MSE_loss:0.09262 epoch [1044/3001], MSE_loss:0.09518 epoch [1045/3001], MSE_loss:0.12923 epoch [1046/3001], MSE_loss:0.08285 epoch [1047/3001], MSE_loss:0.10005 epoch [1048/3001], MSE_loss:0.12070 epoch [1049/3001], MSE_loss:0.11625 epoch [1050/3001], MSE_loss:0.09859 epoch [1051/3001], MSE_loss:0.08482 epoch [1052/3001], MSE_loss:0.12375 epoch [1053/3001], MSE_loss:0.07207 epoch [1054/3001], MSE_loss:0.10105 epoch [1055/3001], MSE_loss:0.08520 epoch [1056/3001], MSE_loss:0.13475 epoch [1057/3001], MSE_loss:0.10670 epoch [1058/3001], MSE_loss:0.12680 epoch [1059/3001], MSE_loss:0.13077 epoch [1060/3001], MSE_loss:0.08601 epoch [1061/3001], MSE_loss:0.09836 epoch [1062/3001], MSE_loss:0.07177 epoch [1063/3001], MSE_loss:0.07671 epoch [1064/3001], MSE_loss:0.12273 epoch [1065/3001], MSE_loss:0.10778 epoch [1066/3001], MSE_loss:0.09331 epoch [1067/3001], MSE_loss:0.10509 epoch [1068/3001], MSE_loss:0.08868 epoch [1069/3001], MSE_loss:0.07949 epoch [1070/3001], MSE_loss:0.05683 epoch [1071/3001], MSE_loss:0.08297 epoch [1072/3001], MSE_loss:0.07440 epoch [1073/3001], MSE_loss:0.07317 epoch [1074/3001], MSE_loss:0.08643 epoch [1075/3001], MSE_loss:0.08960 epoch [1076/3001], MSE_loss:0.10594 epoch [1077/3001], MSE_loss:0.11304 epoch [1078/3001], MSE_loss:0.09280 epoch [1079/3001], MSE_loss:0.09455 epoch [1080/3001], MSE_loss:0.10640 epoch [1081/3001], MSE_loss:0.09696 epoch [1082/3001], MSE_loss:0.11299 epoch [1083/3001], MSE_loss:0.12435 epoch [1084/3001], MSE_loss:0.11955 epoch [1085/3001], MSE_loss:0.11038 epoch [1086/3001], MSE_loss:0.11890 epoch [1087/3001], MSE_loss:0.12079 epoch [1088/3001], MSE_loss:0.09689 epoch [1089/3001], MSE_loss:0.10232 epoch [1090/3001], MSE_loss:0.13510 epoch [1091/3001], MSE_loss:0.10200 epoch [1092/3001], MSE_loss:0.07413 epoch [1093/3001], MSE_loss:0.08392 epoch [1094/3001], MSE_loss:0.08564 epoch [1095/3001], MSE_loss:0.09453 epoch [1096/3001], MSE_loss:0.08954 epoch [1097/3001], MSE_loss:0.10306 epoch [1098/3001], MSE_loss:0.09185 epoch [1099/3001], MSE_loss:0.08546 epoch [1100/3001], MSE_loss:0.07989 epoch [1101/3001], MSE_loss:0.07156 epoch [1102/3001], MSE_loss:0.07831 epoch [1103/3001], MSE_loss:0.15362 epoch [1104/3001], MSE_loss:0.07494 epoch [1105/3001], MSE_loss:0.07838 epoch [1106/3001], MSE_loss:0.08811 epoch [1107/3001], MSE_loss:0.12048 epoch [1108/3001], MSE_loss:0.08952 epoch [1109/3001], MSE_loss:0.08615 epoch [1110/3001], MSE_loss:0.08625 epoch [1111/3001], MSE_loss:0.07232 epoch [1112/3001], MSE_loss:0.08620 epoch [1113/3001], MSE_loss:0.13640 epoch [1114/3001], MSE_loss:0.05777 epoch [1115/3001], MSE_loss:0.11636 epoch [1116/3001], MSE_loss:0.10505 epoch [1117/3001], MSE_loss:0.12621 epoch [1118/3001], MSE_loss:0.09994 epoch [1119/3001], MSE_loss:0.10280 epoch [1120/3001], MSE_loss:0.12593 epoch [1121/3001], MSE_loss:0.09621 epoch [1122/3001], MSE_loss:0.09231 epoch [1123/3001], MSE_loss:0.10089 epoch [1124/3001], MSE_loss:0.09240 epoch [1125/3001], MSE_loss:0.08882 epoch [1126/3001], MSE_loss:0.09348 epoch [1127/3001], MSE_loss:0.10490 epoch [1128/3001], MSE_loss:0.12660 epoch [1129/3001], MSE_loss:0.11345 epoch [1130/3001], MSE_loss:0.13171 epoch [1131/3001], MSE_loss:0.10992 epoch [1132/3001], MSE_loss:0.09241 epoch [1133/3001], MSE_loss:0.10857 epoch [1134/3001], MSE_loss:0.09311 epoch [1135/3001], MSE_loss:0.08096 epoch [1136/3001], MSE_loss:0.10547 epoch [1137/3001], MSE_loss:0.09234 epoch [1138/3001], MSE_loss:0.06472 epoch [1139/3001], MSE_loss:0.09114 epoch [1140/3001], MSE_loss:0.08739 epoch [1141/3001], MSE_loss:0.09323 epoch [1142/3001], MSE_loss:0.07153 epoch [1143/3001], MSE_loss:0.08157 epoch [1144/3001], MSE_loss:0.09792 epoch [1145/3001], MSE_loss:0.07912 epoch [1146/3001], MSE_loss:0.09295 epoch [1147/3001], MSE_loss:0.06906 epoch [1148/3001], MSE_loss:0.12307 epoch [1149/3001], MSE_loss:0.07913 epoch [1150/3001], MSE_loss:0.07991 epoch [1151/3001], MSE_loss:0.11365 epoch [1152/3001], MSE_loss:0.18133 epoch [1153/3001], MSE_loss:0.10253 epoch [1154/3001], MSE_loss:0.07527 epoch [1155/3001], MSE_loss:0.07315 epoch [1156/3001], MSE_loss:0.12350 epoch [1157/3001], MSE_loss:0.08146 epoch [1158/3001], MSE_loss:0.14674 epoch [1159/3001], MSE_loss:0.09566 epoch [1160/3001], MSE_loss:0.07506 epoch [1161/3001], MSE_loss:0.06816 epoch [1162/3001], MSE_loss:0.09486 epoch [1163/3001], MSE_loss:0.11056 epoch [1164/3001], MSE_loss:0.08015 epoch [1165/3001], MSE_loss:0.07345 epoch [1166/3001], MSE_loss:0.11829 epoch [1167/3001], MSE_loss:0.10842 epoch [1168/3001], MSE_loss:0.07547 epoch [1169/3001], MSE_loss:0.10134 epoch [1170/3001], MSE_loss:0.09168 epoch [1171/3001], MSE_loss:0.09484 epoch [1172/3001], MSE_loss:0.08768 epoch [1173/3001], MSE_loss:0.11074 epoch [1174/3001], MSE_loss:0.07795 epoch [1175/3001], MSE_loss:0.08723 epoch [1176/3001], MSE_loss:0.11409 epoch [1177/3001], MSE_loss:0.11711 epoch [1178/3001], MSE_loss:0.12232 epoch [1179/3001], MSE_loss:0.08510 epoch [1180/3001], MSE_loss:0.12254 epoch [1181/3001], MSE_loss:0.09370 epoch [1182/3001], MSE_loss:0.12010 epoch [1183/3001], MSE_loss:0.09975 epoch [1184/3001], MSE_loss:0.11159 epoch [1185/3001], MSE_loss:0.13554 epoch [1186/3001], MSE_loss:0.07492 epoch [1187/3001], MSE_loss:0.11762 epoch [1188/3001], MSE_loss:0.10522 epoch [1189/3001], MSE_loss:0.09603 epoch [1190/3001], MSE_loss:0.11114 epoch [1191/3001], MSE_loss:0.10181 epoch [1192/3001], MSE_loss:0.11511 epoch [1193/3001], MSE_loss:0.10192 epoch [1194/3001], MSE_loss:0.06350 epoch [1195/3001], MSE_loss:0.10962 epoch [1196/3001], MSE_loss:0.07811 epoch [1197/3001], MSE_loss:0.09659 epoch [1198/3001], MSE_loss:0.06888 epoch [1199/3001], MSE_loss:0.08595 epoch [1200/3001], MSE_loss:0.10506 epoch [1201/3001], MSE_loss:0.04808 epoch [1202/3001], MSE_loss:0.08671 epoch [1203/3001], MSE_loss:0.07882 epoch [1204/3001], MSE_loss:0.10200 epoch [1205/3001], MSE_loss:0.08810 epoch [1206/3001], MSE_loss:0.08914 epoch [1207/3001], MSE_loss:0.08367 epoch [1208/3001], MSE_loss:0.09376 epoch [1209/3001], MSE_loss:0.06042 epoch [1210/3001], MSE_loss:0.08429 epoch [1211/3001], MSE_loss:0.08177 epoch [1212/3001], MSE_loss:0.12269 epoch [1213/3001], MSE_loss:0.12817 epoch [1214/3001], MSE_loss:0.09267 epoch [1215/3001], MSE_loss:0.11297 epoch [1216/3001], MSE_loss:0.09496 epoch [1217/3001], MSE_loss:0.11741 epoch [1218/3001], MSE_loss:0.06648 epoch [1219/3001], MSE_loss:0.11088 epoch [1220/3001], MSE_loss:0.08779 epoch [1221/3001], MSE_loss:0.09986 epoch [1222/3001], MSE_loss:0.07972 epoch [1223/3001], MSE_loss:0.07735 epoch [1224/3001], MSE_loss:0.09670 epoch [1225/3001], MSE_loss:0.09078 epoch [1226/3001], MSE_loss:0.07806 epoch [1227/3001], MSE_loss:0.09758 epoch [1228/3001], MSE_loss:0.10429 epoch [1229/3001], MSE_loss:0.07226 epoch [1230/3001], MSE_loss:0.10848 epoch [1231/3001], MSE_loss:0.12376 epoch [1232/3001], MSE_loss:0.05935 epoch [1233/3001], MSE_loss:0.09401 epoch [1234/3001], MSE_loss:0.07550 epoch [1235/3001], MSE_loss:0.11832 epoch [1236/3001], MSE_loss:0.10807 epoch [1237/3001], MSE_loss:0.13853 epoch [1238/3001], MSE_loss:0.10381 epoch [1239/3001], MSE_loss:0.08919 epoch [1240/3001], MSE_loss:0.09345 epoch [1241/3001], MSE_loss:0.09953 epoch [1242/3001], MSE_loss:0.08507 epoch [1243/3001], MSE_loss:0.10868 epoch [1244/3001], MSE_loss:0.09052 epoch [1245/3001], MSE_loss:0.09988 epoch [1246/3001], MSE_loss:0.09028 epoch [1247/3001], MSE_loss:0.09378 epoch [1248/3001], MSE_loss:0.14831 epoch [1249/3001], MSE_loss:0.06205 epoch [1250/3001], MSE_loss:0.08322 epoch [1251/3001], MSE_loss:0.10003 epoch [1252/3001], MSE_loss:0.14299 epoch [1253/3001], MSE_loss:0.10044 epoch [1254/3001], MSE_loss:0.11130 epoch [1255/3001], MSE_loss:0.13665 epoch [1256/3001], MSE_loss:0.10989 epoch [1257/3001], MSE_loss:0.12386 epoch [1258/3001], MSE_loss:0.07107 epoch [1259/3001], MSE_loss:0.08826 epoch [1260/3001], MSE_loss:0.11599 epoch [1261/3001], MSE_loss:0.09512 epoch [1262/3001], MSE_loss:0.06021 epoch [1263/3001], MSE_loss:0.09158 epoch [1264/3001], MSE_loss:0.14700 epoch [1265/3001], MSE_loss:0.11089 epoch [1266/3001], MSE_loss:0.09327 epoch [1267/3001], MSE_loss:0.08106 epoch [1268/3001], MSE_loss:0.10072 epoch [1269/3001], MSE_loss:0.07741 epoch [1270/3001], MSE_loss:0.13507 epoch [1271/3001], MSE_loss:0.12652 epoch [1272/3001], MSE_loss:0.11281 epoch [1273/3001], MSE_loss:0.08147 epoch [1274/3001], MSE_loss:0.10318 epoch [1275/3001], MSE_loss:0.10687 epoch [1276/3001], MSE_loss:0.08000 epoch [1277/3001], MSE_loss:0.09902 epoch [1278/3001], MSE_loss:0.07447 epoch [1279/3001], MSE_loss:0.08596 epoch [1280/3001], MSE_loss:0.08816 epoch [1281/3001], MSE_loss:0.11528 epoch [1282/3001], MSE_loss:0.07632 epoch [1283/3001], MSE_loss:0.09592 epoch [1284/3001], MSE_loss:0.07985 epoch [1285/3001], MSE_loss:0.09572 epoch [1286/3001], MSE_loss:0.08894 epoch [1287/3001], MSE_loss:0.07508 epoch [1288/3001], MSE_loss:0.08676 epoch [1289/3001], MSE_loss:0.10819 epoch [1290/3001], MSE_loss:0.07048 epoch [1291/3001], MSE_loss:0.06653 epoch [1292/3001], MSE_loss:0.11656 epoch [1293/3001], MSE_loss:0.06888 epoch [1294/3001], MSE_loss:0.07819 epoch [1295/3001], MSE_loss:0.09189 epoch [1296/3001], MSE_loss:0.08865 epoch [1297/3001], MSE_loss:0.10418 epoch [1298/3001], MSE_loss:0.08837 epoch [1299/3001], MSE_loss:0.11249 epoch [1300/3001], MSE_loss:0.08009 epoch [1301/3001], MSE_loss:0.10369 epoch [1302/3001], MSE_loss:0.09049 epoch [1303/3001], MSE_loss:0.14753 epoch [1304/3001], MSE_loss:0.06470 epoch [1305/3001], MSE_loss:0.09294 epoch [1306/3001], MSE_loss:0.12387 epoch [1307/3001], MSE_loss:0.09602 epoch [1308/3001], MSE_loss:0.16051 epoch [1309/3001], MSE_loss:0.11153 epoch [1310/3001], MSE_loss:0.09548 epoch [1311/3001], MSE_loss:0.11249 epoch [1312/3001], MSE_loss:0.08649 epoch [1313/3001], MSE_loss:0.10997 epoch [1314/3001], MSE_loss:0.10598 epoch [1315/3001], MSE_loss:0.10788 epoch [1316/3001], MSE_loss:0.05922 epoch [1317/3001], MSE_loss:0.11796 epoch [1318/3001], MSE_loss:0.08587 epoch [1319/3001], MSE_loss:0.08820 epoch [1320/3001], MSE_loss:0.06977 epoch [1321/3001], MSE_loss:0.09997 epoch [1322/3001], MSE_loss:0.11911 epoch [1323/3001], MSE_loss:0.08181 epoch [1324/3001], MSE_loss:0.12025 epoch [1325/3001], MSE_loss:0.07327 epoch [1326/3001], MSE_loss:0.09508 epoch [1327/3001], MSE_loss:0.08901 epoch [1328/3001], MSE_loss:0.07537 epoch [1329/3001], MSE_loss:0.08852 epoch [1330/3001], MSE_loss:0.06650 epoch [1331/3001], MSE_loss:0.13527 epoch [1332/3001], MSE_loss:0.11471 epoch [1333/3001], MSE_loss:0.11441 epoch [1334/3001], MSE_loss:0.12278 epoch [1335/3001], MSE_loss:0.08861 epoch [1336/3001], MSE_loss:0.07168 epoch [1337/3001], MSE_loss:0.11027 epoch [1338/3001], MSE_loss:0.09821 epoch [1339/3001], MSE_loss:0.09501 epoch [1340/3001], MSE_loss:0.07932 epoch [1341/3001], MSE_loss:0.08315 epoch [1342/3001], MSE_loss:0.12695 epoch [1343/3001], MSE_loss:0.08972 epoch [1344/3001], MSE_loss:0.08271 epoch [1345/3001], MSE_loss:0.11742 epoch [1346/3001], MSE_loss:0.08120 epoch [1347/3001], MSE_loss:0.08695 epoch [1348/3001], MSE_loss:0.08686 epoch [1349/3001], MSE_loss:0.08297 epoch [1350/3001], MSE_loss:0.10592 epoch [1351/3001], MSE_loss:0.11898 epoch [1352/3001], MSE_loss:0.08280 epoch [1353/3001], MSE_loss:0.07344 epoch [1354/3001], MSE_loss:0.11591 epoch [1355/3001], MSE_loss:0.10140 epoch [1356/3001], MSE_loss:0.13099 epoch [1357/3001], MSE_loss:0.11233 epoch [1358/3001], MSE_loss:0.13234 epoch [1359/3001], MSE_loss:0.12679 epoch [1360/3001], MSE_loss:0.09450 epoch [1361/3001], MSE_loss:0.08121 epoch [1362/3001], MSE_loss:0.07970 epoch [1363/3001], MSE_loss:0.06691 epoch [1364/3001], MSE_loss:0.10510 epoch [1365/3001], MSE_loss:0.10833 epoch [1366/3001], MSE_loss:0.08854 epoch [1367/3001], MSE_loss:0.08013 epoch [1368/3001], MSE_loss:0.08701 epoch [1369/3001], MSE_loss:0.10064 epoch [1370/3001], MSE_loss:0.14501 epoch [1371/3001], MSE_loss:0.10182 epoch [1372/3001], MSE_loss:0.10489 epoch [1373/3001], MSE_loss:0.11309 epoch [1374/3001], MSE_loss:0.09379 epoch [1375/3001], MSE_loss:0.08237 epoch [1376/3001], MSE_loss:0.10782 epoch [1377/3001], MSE_loss:0.10002 epoch [1378/3001], MSE_loss:0.06435 epoch [1379/3001], MSE_loss:0.15944 epoch [1380/3001], MSE_loss:0.12420 epoch [1381/3001], MSE_loss:0.07682 epoch [1382/3001], MSE_loss:0.05946 epoch [1383/3001], MSE_loss:0.11447 epoch [1384/3001], MSE_loss:0.07097 epoch [1385/3001], MSE_loss:0.08505 epoch [1386/3001], MSE_loss:0.12295 epoch [1387/3001], MSE_loss:0.08855 epoch [1388/3001], MSE_loss:0.15208 epoch [1389/3001], MSE_loss:0.08334 epoch [1390/3001], MSE_loss:0.07880 epoch [1391/3001], MSE_loss:0.09002 epoch [1392/3001], MSE_loss:0.09091 epoch [1393/3001], MSE_loss:0.09965 epoch [1394/3001], MSE_loss:0.10199 epoch [1395/3001], MSE_loss:0.10111 epoch [1396/3001], MSE_loss:0.09290 epoch [1397/3001], MSE_loss:0.08639 epoch [1398/3001], MSE_loss:0.10021 epoch [1399/3001], MSE_loss:0.09299 epoch [1400/3001], MSE_loss:0.14173 epoch [1401/3001], MSE_loss:0.08867 epoch [1402/3001], MSE_loss:0.10930 epoch [1403/3001], MSE_loss:0.09208 epoch [1404/3001], MSE_loss:0.12390 epoch [1405/3001], MSE_loss:0.08633 epoch [1406/3001], MSE_loss:0.08591 epoch [1407/3001], MSE_loss:0.12402 epoch [1408/3001], MSE_loss:0.08290 epoch [1409/3001], MSE_loss:0.07237 epoch [1410/3001], MSE_loss:0.08248 epoch [1411/3001], MSE_loss:0.11674 epoch [1412/3001], MSE_loss:0.09069 epoch [1413/3001], MSE_loss:0.07753 epoch [1414/3001], MSE_loss:0.12553 epoch [1415/3001], MSE_loss:0.08203 epoch [1416/3001], MSE_loss:0.09919 epoch [1417/3001], MSE_loss:0.08632 epoch [1418/3001], MSE_loss:0.09924 epoch [1419/3001], MSE_loss:0.09967 epoch [1420/3001], MSE_loss:0.09472 epoch [1421/3001], MSE_loss:0.10588 epoch [1422/3001], MSE_loss:0.09018 epoch [1423/3001], MSE_loss:0.08626 epoch [1424/3001], MSE_loss:0.11087 epoch [1425/3001], MSE_loss:0.07695 epoch [1426/3001], MSE_loss:0.07963 epoch [1427/3001], MSE_loss:0.13196 epoch [1428/3001], MSE_loss:0.09758 epoch [1429/3001], MSE_loss:0.08797 epoch [1430/3001], MSE_loss:0.10740 epoch [1431/3001], MSE_loss:0.06405 epoch [1432/3001], MSE_loss:0.09635 epoch [1433/3001], MSE_loss:0.06882 epoch [1434/3001], MSE_loss:0.08710 epoch [1435/3001], MSE_loss:0.07739 epoch [1436/3001], MSE_loss:0.11263 epoch [1437/3001], MSE_loss:0.08626 epoch [1438/3001], MSE_loss:0.08875 epoch [1439/3001], MSE_loss:0.08662 epoch [1440/3001], MSE_loss:0.10686 epoch [1441/3001], MSE_loss:0.07724 epoch [1442/3001], MSE_loss:0.11008 epoch [1443/3001], MSE_loss:0.08608 epoch [1444/3001], MSE_loss:0.07972 epoch [1445/3001], MSE_loss:0.08290 epoch [1446/3001], MSE_loss:0.11938 epoch [1447/3001], MSE_loss:0.06751 epoch [1448/3001], MSE_loss:0.07312 epoch [1449/3001], MSE_loss:0.08875 epoch [1450/3001], MSE_loss:0.08580 epoch [1451/3001], MSE_loss:0.08352 epoch [1452/3001], MSE_loss:0.08540 epoch [1453/3001], MSE_loss:0.09322 epoch [1454/3001], MSE_loss:0.06870 epoch [1455/3001], MSE_loss:0.08588 epoch [1456/3001], MSE_loss:0.12793 epoch [1457/3001], MSE_loss:0.08353 epoch [1458/3001], MSE_loss:0.09567 epoch [1459/3001], MSE_loss:0.08839 epoch [1460/3001], MSE_loss:0.08672 epoch [1461/3001], MSE_loss:0.11075 epoch [1462/3001], MSE_loss:0.07310 epoch [1463/3001], MSE_loss:0.10747 epoch [1464/3001], MSE_loss:0.10448 epoch [1465/3001], MSE_loss:0.10316 epoch [1466/3001], MSE_loss:0.11977 epoch [1467/3001], MSE_loss:0.07302 epoch [1468/3001], MSE_loss:0.11956 epoch [1469/3001], MSE_loss:0.07535 epoch [1470/3001], MSE_loss:0.08605 epoch [1471/3001], MSE_loss:0.09748 epoch [1472/3001], MSE_loss:0.08699 epoch [1473/3001], MSE_loss:0.09901 epoch [1474/3001], MSE_loss:0.08552 epoch [1475/3001], MSE_loss:0.09016 epoch [1476/3001], MSE_loss:0.08694 epoch [1477/3001], MSE_loss:0.08079 epoch [1478/3001], MSE_loss:0.07339 epoch [1479/3001], MSE_loss:0.12158 epoch [1480/3001], MSE_loss:0.08186 epoch [1481/3001], MSE_loss:0.07575 epoch [1482/3001], MSE_loss:0.08380 epoch [1483/3001], MSE_loss:0.08220 epoch [1484/3001], MSE_loss:0.12267 epoch [1485/3001], MSE_loss:0.14257 epoch [1486/3001], MSE_loss:0.07948 epoch [1487/3001], MSE_loss:0.11429 epoch [1488/3001], MSE_loss:0.12453 epoch [1489/3001], MSE_loss:0.06808 epoch [1490/3001], MSE_loss:0.12820 epoch [1491/3001], MSE_loss:0.10374 epoch [1492/3001], MSE_loss:0.07415 epoch [1493/3001], MSE_loss:0.08180 epoch [1494/3001], MSE_loss:0.10006 epoch [1495/3001], MSE_loss:0.06376 epoch [1496/3001], MSE_loss:0.08953 epoch [1497/3001], MSE_loss:0.06038 epoch [1498/3001], MSE_loss:0.13776 epoch [1499/3001], MSE_loss:0.06590 epoch [1500/3001], MSE_loss:0.08654 epoch [1501/3001], MSE_loss:0.08548 epoch [1502/3001], MSE_loss:0.11364 epoch [1503/3001], MSE_loss:0.07023 epoch [1504/3001], MSE_loss:0.09002 epoch [1505/3001], MSE_loss:0.10250 epoch [1506/3001], MSE_loss:0.06662 epoch [1507/3001], MSE_loss:0.09575 epoch [1508/3001], MSE_loss:0.07635 epoch [1509/3001], MSE_loss:0.06084 epoch [1510/3001], MSE_loss:0.16553 epoch [1511/3001], MSE_loss:0.10348 epoch [1512/3001], MSE_loss:0.06953 epoch [1513/3001], MSE_loss:0.11350 epoch [1514/3001], MSE_loss:0.09145 epoch [1515/3001], MSE_loss:0.09935 epoch [1516/3001], MSE_loss:0.07958 epoch [1517/3001], MSE_loss:0.11643 epoch [1518/3001], MSE_loss:0.09896 epoch [1519/3001], MSE_loss:0.09388 epoch [1520/3001], MSE_loss:0.09117 epoch [1521/3001], MSE_loss:0.07967 epoch [1522/3001], MSE_loss:0.06807 epoch [1523/3001], MSE_loss:0.13979 epoch [1524/3001], MSE_loss:0.09916 epoch [1525/3001], MSE_loss:0.08597 epoch [1526/3001], MSE_loss:0.06660 epoch [1527/3001], MSE_loss:0.08162 epoch [1528/3001], MSE_loss:0.11477 epoch [1529/3001], MSE_loss:0.08462 epoch [1530/3001], MSE_loss:0.07294 epoch [1531/3001], MSE_loss:0.06198 epoch [1532/3001], MSE_loss:0.08420 epoch [1533/3001], MSE_loss:0.11220 epoch [1534/3001], MSE_loss:0.08870 epoch [1535/3001], MSE_loss:0.09642 epoch [1536/3001], MSE_loss:0.06970 epoch [1537/3001], MSE_loss:0.08269 epoch [1538/3001], MSE_loss:0.11841 epoch [1539/3001], MSE_loss:0.06626 epoch [1540/3001], MSE_loss:0.08488 epoch [1541/3001], MSE_loss:0.09114 epoch [1542/3001], MSE_loss:0.10366 epoch [1543/3001], MSE_loss:0.06702 epoch [1544/3001], MSE_loss:0.09357 epoch [1545/3001], MSE_loss:0.08744 epoch [1546/3001], MSE_loss:0.12382 epoch [1547/3001], MSE_loss:0.08724 epoch [1548/3001], MSE_loss:0.06735 epoch [1549/3001], MSE_loss:0.09450 epoch [1550/3001], MSE_loss:0.10736 epoch [1551/3001], MSE_loss:0.07984 epoch [1552/3001], MSE_loss:0.07543 epoch [1553/3001], MSE_loss:0.06111 epoch [1554/3001], MSE_loss:0.10124 epoch [1555/3001], MSE_loss:0.07390 epoch [1556/3001], MSE_loss:0.09949 epoch [1557/3001], MSE_loss:0.06072 epoch [1558/3001], MSE_loss:0.06903 epoch [1559/3001], MSE_loss:0.08474 epoch [1560/3001], MSE_loss:0.09944 epoch [1561/3001], MSE_loss:0.08395 epoch [1562/3001], MSE_loss:0.08817 epoch [1563/3001], MSE_loss:0.13487 epoch [1564/3001], MSE_loss:0.10166 epoch [1565/3001], MSE_loss:0.06624 epoch [1566/3001], MSE_loss:0.08525 epoch [1567/3001], MSE_loss:0.10649 epoch [1568/3001], MSE_loss:0.08321 epoch [1569/3001], MSE_loss:0.09481 epoch [1570/3001], MSE_loss:0.10903 epoch [1571/3001], MSE_loss:0.08892 epoch [1572/3001], MSE_loss:0.08699 epoch [1573/3001], MSE_loss:0.09431 epoch [1574/3001], MSE_loss:0.08354 epoch [1575/3001], MSE_loss:0.08426 epoch [1576/3001], MSE_loss:0.08419 epoch [1577/3001], MSE_loss:0.10038 epoch [1578/3001], MSE_loss:0.08839 epoch [1579/3001], MSE_loss:0.07550 epoch [1580/3001], MSE_loss:0.09263 epoch [1581/3001], MSE_loss:0.09773 epoch [1582/3001], MSE_loss:0.11309 epoch [1583/3001], MSE_loss:0.07738 epoch [1584/3001], MSE_loss:0.11252 epoch [1585/3001], MSE_loss:0.09868 epoch [1586/3001], MSE_loss:0.08722 epoch [1587/3001], MSE_loss:0.11144 epoch [1588/3001], MSE_loss:0.07481 epoch [1589/3001], MSE_loss:0.10438 epoch [1590/3001], MSE_loss:0.11053 epoch [1591/3001], MSE_loss:0.10062 epoch [1592/3001], MSE_loss:0.11736 epoch [1593/3001], MSE_loss:0.08863 epoch [1594/3001], MSE_loss:0.07112 epoch [1595/3001], MSE_loss:0.11175 epoch [1596/3001], MSE_loss:0.05500 epoch [1597/3001], MSE_loss:0.09079 epoch [1598/3001], MSE_loss:0.08520 epoch [1599/3001], MSE_loss:0.09357 epoch [1600/3001], MSE_loss:0.12201 epoch [1601/3001], MSE_loss:0.10329 epoch [1602/3001], MSE_loss:0.10004 epoch [1603/3001], MSE_loss:0.14520 epoch [1604/3001], MSE_loss:0.10245 epoch [1605/3001], MSE_loss:0.07824 epoch [1606/3001], MSE_loss:0.10258 epoch [1607/3001], MSE_loss:0.10076 epoch [1608/3001], MSE_loss:0.13165 epoch [1609/3001], MSE_loss:0.08795 epoch [1610/3001], MSE_loss:0.08915 epoch [1611/3001], MSE_loss:0.10185 epoch [1612/3001], MSE_loss:0.08111 epoch [1613/3001], MSE_loss:0.12522 epoch [1614/3001], MSE_loss:0.10742 epoch [1615/3001], MSE_loss:0.10416 epoch [1616/3001], MSE_loss:0.10315 epoch [1617/3001], MSE_loss:0.10738 epoch [1618/3001], MSE_loss:0.07198 epoch [1619/3001], MSE_loss:0.09478 epoch [1620/3001], MSE_loss:0.09897 epoch [1621/3001], MSE_loss:0.12421 epoch [1622/3001], MSE_loss:0.08078 epoch [1623/3001], MSE_loss:0.08831 epoch [1624/3001], MSE_loss:0.05857 epoch [1625/3001], MSE_loss:0.11711 epoch [1626/3001], MSE_loss:0.10121 epoch [1627/3001], MSE_loss:0.09760 epoch [1628/3001], MSE_loss:0.09811 epoch [1629/3001], MSE_loss:0.10600 epoch [1630/3001], MSE_loss:0.09215 epoch [1631/3001], MSE_loss:0.05336 epoch [1632/3001], MSE_loss:0.16741 epoch [1633/3001], MSE_loss:0.08373 epoch [1634/3001], MSE_loss:0.07409 epoch [1635/3001], MSE_loss:0.10090 epoch [1636/3001], MSE_loss:0.08718 epoch [1637/3001], MSE_loss:0.12598 epoch [1638/3001], MSE_loss:0.07549 epoch [1639/3001], MSE_loss:0.06748 epoch [1640/3001], MSE_loss:0.09920 epoch [1641/3001], MSE_loss:0.06349 epoch [1642/3001], MSE_loss:0.12648 epoch [1643/3001], MSE_loss:0.08230 epoch [1644/3001], MSE_loss:0.10197 epoch [1645/3001], MSE_loss:0.07202 epoch [1646/3001], MSE_loss:0.10555 epoch [1647/3001], MSE_loss:0.09679 epoch [1648/3001], MSE_loss:0.09550 epoch [1649/3001], MSE_loss:0.06579 epoch [1650/3001], MSE_loss:0.05432 epoch [1651/3001], MSE_loss:0.06943 epoch [1652/3001], MSE_loss:0.08664 epoch [1653/3001], MSE_loss:0.11342 epoch [1654/3001], MSE_loss:0.11287 epoch [1655/3001], MSE_loss:0.10801 epoch [1656/3001], MSE_loss:0.09101 epoch [1657/3001], MSE_loss:0.09153 epoch [1658/3001], MSE_loss:0.11418 epoch [1659/3001], MSE_loss:0.10570 epoch [1660/3001], MSE_loss:0.12087 epoch [1661/3001], MSE_loss:0.08708 epoch [1662/3001], MSE_loss:0.11036 epoch [1663/3001], MSE_loss:0.09464 epoch [1664/3001], MSE_loss:0.08038 epoch [1665/3001], MSE_loss:0.13033 epoch [1666/3001], MSE_loss:0.08915 epoch [1667/3001], MSE_loss:0.12211 epoch [1668/3001], MSE_loss:0.12376 epoch [1669/3001], MSE_loss:0.14226 epoch [1670/3001], MSE_loss:0.09887 epoch [1671/3001], MSE_loss:0.06802 epoch [1672/3001], MSE_loss:0.11053 epoch [1673/3001], MSE_loss:0.08180 epoch [1674/3001], MSE_loss:0.06937 epoch [1675/3001], MSE_loss:0.08802 epoch [1676/3001], MSE_loss:0.11411 epoch [1677/3001], MSE_loss:0.09817 epoch [1678/3001], MSE_loss:0.11250 epoch [1679/3001], MSE_loss:0.10042 epoch [1680/3001], MSE_loss:0.10645 epoch [1681/3001], MSE_loss:0.08513 epoch [1682/3001], MSE_loss:0.08015 epoch [1683/3001], MSE_loss:0.11263 epoch [1684/3001], MSE_loss:0.13175 epoch [1685/3001], MSE_loss:0.08800 epoch [1686/3001], MSE_loss:0.08088 epoch [1687/3001], MSE_loss:0.08683 epoch [1688/3001], MSE_loss:0.09713 epoch [1689/3001], MSE_loss:0.10977 epoch [1690/3001], MSE_loss:0.07923 epoch [1691/3001], MSE_loss:0.09380 epoch [1692/3001], MSE_loss:0.09443 epoch [1693/3001], MSE_loss:0.08743 epoch [1694/3001], MSE_loss:0.08611 epoch [1695/3001], MSE_loss:0.05664 epoch [1696/3001], MSE_loss:0.07098 epoch [1697/3001], MSE_loss:0.10221 epoch [1698/3001], MSE_loss:0.12660 epoch [1699/3001], MSE_loss:0.09991 epoch [1700/3001], MSE_loss:0.10124 epoch [1701/3001], MSE_loss:0.10681 epoch [1702/3001], MSE_loss:0.07983 epoch [1703/3001], MSE_loss:0.09368 epoch [1704/3001], MSE_loss:0.10196 epoch [1705/3001], MSE_loss:0.08937 epoch [1706/3001], MSE_loss:0.06131 epoch [1707/3001], MSE_loss:0.08284 epoch [1708/3001], MSE_loss:0.09581 epoch [1709/3001], MSE_loss:0.12272 epoch [1710/3001], MSE_loss:0.07672 epoch [1711/3001], MSE_loss:0.06573 epoch [1712/3001], MSE_loss:0.11751 epoch [1713/3001], MSE_loss:0.07839 epoch [1714/3001], MSE_loss:0.08992 epoch [1715/3001], MSE_loss:0.07169 epoch [1716/3001], MSE_loss:0.07971 epoch [1717/3001], MSE_loss:0.07308 epoch [1718/3001], MSE_loss:0.10365 epoch [1719/3001], MSE_loss:0.08799 epoch [1720/3001], MSE_loss:0.09137 epoch [1721/3001], MSE_loss:0.08279 epoch [1722/3001], MSE_loss:0.07819 epoch [1723/3001], MSE_loss:0.08588 epoch [1724/3001], MSE_loss:0.06646 epoch [1725/3001], MSE_loss:0.13048 epoch [1726/3001], MSE_loss:0.09093 epoch [1727/3001], MSE_loss:0.12126 epoch [1728/3001], MSE_loss:0.10462 epoch [1729/3001], MSE_loss:0.10452 epoch [1730/3001], MSE_loss:0.10972 epoch [1731/3001], MSE_loss:0.08140 epoch [1732/3001], MSE_loss:0.10229 epoch [1733/3001], MSE_loss:0.08367 epoch [1734/3001], MSE_loss:0.10969 epoch [1735/3001], MSE_loss:0.12104 epoch [1736/3001], MSE_loss:0.06487 epoch [1737/3001], MSE_loss:0.09512 epoch [1738/3001], MSE_loss:0.08592 epoch [1739/3001], MSE_loss:0.08327 epoch [1740/3001], MSE_loss:0.11545 epoch [1741/3001], MSE_loss:0.07535 epoch [1742/3001], MSE_loss:0.09999 epoch [1743/3001], MSE_loss:0.09303 epoch [1744/3001], MSE_loss:0.08597 epoch [1745/3001], MSE_loss:0.08289 epoch [1746/3001], MSE_loss:0.11244 epoch [1747/3001], MSE_loss:0.08800 epoch [1748/3001], MSE_loss:0.07510 epoch [1749/3001], MSE_loss:0.11374 epoch [1750/3001], MSE_loss:0.10411 epoch [1751/3001], MSE_loss:0.11730 epoch [1752/3001], MSE_loss:0.09136 epoch [1753/3001], MSE_loss:0.11017 epoch [1754/3001], MSE_loss:0.10498 epoch [1755/3001], MSE_loss:0.06854 epoch [1756/3001], MSE_loss:0.05963 epoch [1757/3001], MSE_loss:0.07995 epoch [1758/3001], MSE_loss:0.13612 epoch [1759/3001], MSE_loss:0.06940 epoch [1760/3001], MSE_loss:0.08479 epoch [1761/3001], MSE_loss:0.07650 epoch [1762/3001], MSE_loss:0.10230 epoch [1763/3001], MSE_loss:0.09167 epoch [1764/3001], MSE_loss:0.06368 epoch [1765/3001], MSE_loss:0.09566 epoch [1766/3001], MSE_loss:0.12590 epoch [1767/3001], MSE_loss:0.07428 epoch [1768/3001], MSE_loss:0.08571 epoch [1769/3001], MSE_loss:0.11204 epoch [1770/3001], MSE_loss:0.08002 epoch [1771/3001], MSE_loss:0.09330 epoch [1772/3001], MSE_loss:0.08576 epoch [1773/3001], MSE_loss:0.12945 epoch [1774/3001], MSE_loss:0.10277 epoch [1775/3001], MSE_loss:0.12774 epoch [1776/3001], MSE_loss:0.09523 epoch [1777/3001], MSE_loss:0.07214 epoch [1778/3001], MSE_loss:0.06615 epoch [1779/3001], MSE_loss:0.07901 epoch [1780/3001], MSE_loss:0.07368 epoch [1781/3001], MSE_loss:0.06396 epoch [1782/3001], MSE_loss:0.10932 epoch [1783/3001], MSE_loss:0.08137 epoch [1784/3001], MSE_loss:0.12553 epoch [1785/3001], MSE_loss:0.07916 epoch [1786/3001], MSE_loss:0.09082 epoch [1787/3001], MSE_loss:0.09629 epoch [1788/3001], MSE_loss:0.07926 epoch [1789/3001], MSE_loss:0.09480 epoch [1790/3001], MSE_loss:0.10262 epoch [1791/3001], MSE_loss:0.11764 epoch [1792/3001], MSE_loss:0.12488 epoch [1793/3001], MSE_loss:0.09426 epoch [1794/3001], MSE_loss:0.07645 epoch [1795/3001], MSE_loss:0.07910 epoch [1796/3001], MSE_loss:0.10173 epoch [1797/3001], MSE_loss:0.09350 epoch [1798/3001], MSE_loss:0.11221 epoch [1799/3001], MSE_loss:0.08311 epoch [1800/3001], MSE_loss:0.09292 epoch [1801/3001], MSE_loss:0.10957 epoch [1802/3001], MSE_loss:0.11604 epoch [1803/3001], MSE_loss:0.07875 epoch [1804/3001], MSE_loss:0.10233 epoch [1805/3001], MSE_loss:0.11784 epoch [1806/3001], MSE_loss:0.05309 epoch [1807/3001], MSE_loss:0.09232 epoch [1808/3001], MSE_loss:0.11687 epoch [1809/3001], MSE_loss:0.09262 epoch [1810/3001], MSE_loss:0.07846 epoch [1811/3001], MSE_loss:0.11029 epoch [1812/3001], MSE_loss:0.08489 epoch [1813/3001], MSE_loss:0.10359 epoch [1814/3001], MSE_loss:0.08692 epoch [1815/3001], MSE_loss:0.09217 epoch [1816/3001], MSE_loss:0.07350 epoch [1817/3001], MSE_loss:0.07835 epoch [1818/3001], MSE_loss:0.08874 epoch [1819/3001], MSE_loss:0.09072 epoch [1820/3001], MSE_loss:0.07705 epoch [1821/3001], MSE_loss:0.16512 epoch [1822/3001], MSE_loss:0.12075 epoch [1823/3001], MSE_loss:0.08708 epoch [1824/3001], MSE_loss:0.09888 epoch [1825/3001], MSE_loss:0.09806 epoch [1826/3001], MSE_loss:0.08863 epoch [1827/3001], MSE_loss:0.06938 epoch [1828/3001], MSE_loss:0.10358 epoch [1829/3001], MSE_loss:0.10969 epoch [1830/3001], MSE_loss:0.08717 epoch [1831/3001], MSE_loss:0.05841 epoch [1832/3001], MSE_loss:0.05954 epoch [1833/3001], MSE_loss:0.11145 epoch [1834/3001], MSE_loss:0.11816 epoch [1835/3001], MSE_loss:0.15713 epoch [1836/3001], MSE_loss:0.08335 epoch [1837/3001], MSE_loss:0.14316 epoch [1838/3001], MSE_loss:0.09156 epoch [1839/3001], MSE_loss:0.10483 epoch [1840/3001], MSE_loss:0.06208 epoch [1841/3001], MSE_loss:0.10412 epoch [1842/3001], MSE_loss:0.09964 epoch [1843/3001], MSE_loss:0.10704 epoch [1844/3001], MSE_loss:0.07354 epoch [1845/3001], MSE_loss:0.13545 epoch [1846/3001], MSE_loss:0.08070 epoch [1847/3001], MSE_loss:0.10415 epoch [1848/3001], MSE_loss:0.09819 epoch [1849/3001], MSE_loss:0.10999 epoch [1850/3001], MSE_loss:0.09917 epoch [1851/3001], MSE_loss:0.08605 epoch [1852/3001], MSE_loss:0.09803 epoch [1853/3001], MSE_loss:0.06131 epoch [1854/3001], MSE_loss:0.08795 epoch [1855/3001], MSE_loss:0.08629 epoch [1856/3001], MSE_loss:0.08366 epoch [1857/3001], MSE_loss:0.11539 epoch [1858/3001], MSE_loss:0.07559 epoch [1859/3001], MSE_loss:0.09036 epoch [1860/3001], MSE_loss:0.09866 epoch [1861/3001], MSE_loss:0.09059 epoch [1862/3001], MSE_loss:0.07990 epoch [1863/3001], MSE_loss:0.07576 epoch [1864/3001], MSE_loss:0.10332 epoch [1865/3001], MSE_loss:0.10852 epoch [1866/3001], MSE_loss:0.12173 epoch [1867/3001], MSE_loss:0.10188 epoch [1868/3001], MSE_loss:0.11243 epoch [1869/3001], MSE_loss:0.07749 epoch [1870/3001], MSE_loss:0.09873 epoch [1871/3001], MSE_loss:0.07222 epoch [1872/3001], MSE_loss:0.08358 epoch [1873/3001], MSE_loss:0.07752 epoch [1874/3001], MSE_loss:0.11773 epoch [1875/3001], MSE_loss:0.09037 epoch [1876/3001], MSE_loss:0.11152 epoch [1877/3001], MSE_loss:0.10478 epoch [1878/3001], MSE_loss:0.14910 epoch [1879/3001], MSE_loss:0.08933 epoch [1880/3001], MSE_loss:0.08136 epoch [1881/3001], MSE_loss:0.11555 epoch [1882/3001], MSE_loss:0.08960 epoch [1883/3001], MSE_loss:0.10589 epoch [1884/3001], MSE_loss:0.11265 epoch [1885/3001], MSE_loss:0.07445 epoch [1886/3001], MSE_loss:0.11144 epoch [1887/3001], MSE_loss:0.09409 epoch [1888/3001], MSE_loss:0.09313 epoch [1889/3001], MSE_loss:0.09749 epoch [1890/3001], MSE_loss:0.10910 epoch [1891/3001], MSE_loss:0.05933 epoch [1892/3001], MSE_loss:0.06455 epoch [1893/3001], MSE_loss:0.08986 epoch [1894/3001], MSE_loss:0.07898 epoch [1895/3001], MSE_loss:0.07253 epoch [1896/3001], MSE_loss:0.07711 epoch [1897/3001], MSE_loss:0.08205 epoch [1898/3001], MSE_loss:0.12264 epoch [1899/3001], MSE_loss:0.12890 epoch [1900/3001], MSE_loss:0.13096 epoch [1901/3001], MSE_loss:0.09234 epoch [1902/3001], MSE_loss:0.08889 epoch [1903/3001], MSE_loss:0.08175 epoch [1904/3001], MSE_loss:0.06564 epoch [1905/3001], MSE_loss:0.09505 epoch [1906/3001], MSE_loss:0.11424 epoch [1907/3001], MSE_loss:0.08592 epoch [1908/3001], MSE_loss:0.06316 epoch [1909/3001], MSE_loss:0.06033 epoch [1910/3001], MSE_loss:0.09209 epoch [1911/3001], MSE_loss:0.13745 epoch [1912/3001], MSE_loss:0.09535 epoch [1913/3001], MSE_loss:0.11697 epoch [1914/3001], MSE_loss:0.11431 epoch [1915/3001], MSE_loss:0.07486 epoch [1916/3001], MSE_loss:0.06441 epoch [1917/3001], MSE_loss:0.10876 epoch [1918/3001], MSE_loss:0.07009 epoch [1919/3001], MSE_loss:0.09473 epoch [1920/3001], MSE_loss:0.07494 epoch [1921/3001], MSE_loss:0.06521 epoch [1922/3001], MSE_loss:0.07945 epoch [1923/3001], MSE_loss:0.09473 epoch [1924/3001], MSE_loss:0.12316 epoch [1925/3001], MSE_loss:0.09148 epoch [1926/3001], MSE_loss:0.11901 epoch [1927/3001], MSE_loss:0.09573 epoch [1928/3001], MSE_loss:0.09727 epoch [1929/3001], MSE_loss:0.10993 epoch [1930/3001], MSE_loss:0.05679 epoch [1931/3001], MSE_loss:0.10817 epoch [1932/3001], MSE_loss:0.07639 epoch [1933/3001], MSE_loss:0.11809 epoch [1934/3001], MSE_loss:0.13112 epoch [1935/3001], MSE_loss:0.09441 epoch [1936/3001], MSE_loss:0.08482 epoch [1937/3001], MSE_loss:0.07496 epoch [1938/3001], MSE_loss:0.09685 epoch [1939/3001], MSE_loss:0.10498 epoch [1940/3001], MSE_loss:0.09775 epoch [1941/3001], MSE_loss:0.07921 epoch [1942/3001], MSE_loss:0.11984 epoch [1943/3001], MSE_loss:0.10018 epoch [1944/3001], MSE_loss:0.09654 epoch [1945/3001], MSE_loss:0.09459 epoch [1946/3001], MSE_loss:0.05812 epoch [1947/3001], MSE_loss:0.10137 epoch [1948/3001], MSE_loss:0.08219 epoch [1949/3001], MSE_loss:0.07265 epoch [1950/3001], MSE_loss:0.08808 epoch [1951/3001], MSE_loss:0.11867 epoch [1952/3001], MSE_loss:0.10133 epoch [1953/3001], MSE_loss:0.08484 epoch [1954/3001], MSE_loss:0.10698 epoch [1955/3001], MSE_loss:0.12585 epoch [1956/3001], MSE_loss:0.11000 epoch [1957/3001], MSE_loss:0.09380 epoch [1958/3001], MSE_loss:0.11079 epoch [1959/3001], MSE_loss:0.07198 epoch [1960/3001], MSE_loss:0.09886 epoch [1961/3001], MSE_loss:0.10553 epoch [1962/3001], MSE_loss:0.10607 epoch [1963/3001], MSE_loss:0.10730 epoch [1964/3001], MSE_loss:0.07713 epoch [1965/3001], MSE_loss:0.14014 epoch [1966/3001], MSE_loss:0.10194 epoch [1967/3001], MSE_loss:0.09819 epoch [1968/3001], MSE_loss:0.10581 epoch [1969/3001], MSE_loss:0.10547 epoch [1970/3001], MSE_loss:0.06490 epoch [1971/3001], MSE_loss:0.08421 epoch [1972/3001], MSE_loss:0.06249 epoch [1973/3001], MSE_loss:0.09583 epoch [1974/3001], MSE_loss:0.08966 epoch [1975/3001], MSE_loss:0.08827 epoch [1976/3001], MSE_loss:0.08052 epoch [1977/3001], MSE_loss:0.12086 epoch [1978/3001], MSE_loss:0.10170 epoch [1979/3001], MSE_loss:0.06935 epoch [1980/3001], MSE_loss:0.09482 epoch [1981/3001], MSE_loss:0.10448 epoch [1982/3001], MSE_loss:0.08002 epoch [1983/3001], MSE_loss:0.07861 epoch [1984/3001], MSE_loss:0.10118 epoch [1985/3001], MSE_loss:0.11918 epoch [1986/3001], MSE_loss:0.08004 epoch [1987/3001], MSE_loss:0.09284 epoch [1988/3001], MSE_loss:0.13807 epoch [1989/3001], MSE_loss:0.08340 epoch [1990/3001], MSE_loss:0.12196 epoch [1991/3001], MSE_loss:0.14275 epoch [1992/3001], MSE_loss:0.06758 epoch [1993/3001], MSE_loss:0.12310 epoch [1994/3001], MSE_loss:0.09548 epoch [1995/3001], MSE_loss:0.08276 epoch [1996/3001], MSE_loss:0.13294 epoch [1997/3001], MSE_loss:0.08150 epoch [1998/3001], MSE_loss:0.12992 epoch [1999/3001], MSE_loss:0.07631 epoch [2000/3001], MSE_loss:0.08671 epoch [2001/3001], MSE_loss:0.10335 epoch [2002/3001], MSE_loss:0.08625 epoch [2003/3001], MSE_loss:0.09948 epoch [2004/3001], MSE_loss:0.09644 epoch [2005/3001], MSE_loss:0.11481 epoch [2006/3001], MSE_loss:0.09620 epoch [2007/3001], MSE_loss:0.11308 epoch [2008/3001], MSE_loss:0.10668 epoch [2009/3001], MSE_loss:0.07075 epoch [2010/3001], MSE_loss:0.08049 epoch [2011/3001], MSE_loss:0.06880 epoch [2012/3001], MSE_loss:0.08166 epoch [2013/3001], MSE_loss:0.09622 epoch [2014/3001], MSE_loss:0.07323 epoch [2015/3001], MSE_loss:0.09198 epoch [2016/3001], MSE_loss:0.09041 epoch [2017/3001], MSE_loss:0.08389 epoch [2018/3001], MSE_loss:0.06996 epoch [2019/3001], MSE_loss:0.08875 epoch [2020/3001], MSE_loss:0.09776 epoch [2021/3001], MSE_loss:0.12070 epoch [2022/3001], MSE_loss:0.08991 epoch [2023/3001], MSE_loss:0.14300 epoch [2024/3001], MSE_loss:0.10740 epoch [2025/3001], MSE_loss:0.12002 epoch [2026/3001], MSE_loss:0.13303 epoch [2027/3001], MSE_loss:0.06409 epoch [2028/3001], MSE_loss:0.07590 epoch [2029/3001], MSE_loss:0.16658 epoch [2030/3001], MSE_loss:0.06638 epoch [2031/3001], MSE_loss:0.10147 epoch [2032/3001], MSE_loss:0.08759 epoch [2033/3001], MSE_loss:0.09005 epoch [2034/3001], MSE_loss:0.09042 epoch [2035/3001], MSE_loss:0.12312 epoch [2036/3001], MSE_loss:0.07355 epoch [2037/3001], MSE_loss:0.11138 epoch [2038/3001], MSE_loss:0.08637 epoch [2039/3001], MSE_loss:0.09276 epoch [2040/3001], MSE_loss:0.08439 epoch [2041/3001], MSE_loss:0.09260 epoch [2042/3001], MSE_loss:0.10862 epoch [2043/3001], MSE_loss:0.08860 epoch [2044/3001], MSE_loss:0.07624 epoch [2045/3001], MSE_loss:0.12241 epoch [2046/3001], MSE_loss:0.10382 epoch [2047/3001], MSE_loss:0.07236 epoch [2048/3001], MSE_loss:0.05742 epoch [2049/3001], MSE_loss:0.10438 epoch [2050/3001], MSE_loss:0.10356 epoch [2051/3001], MSE_loss:0.08121 epoch [2052/3001], MSE_loss:0.08176 epoch [2053/3001], MSE_loss:0.08349 epoch [2054/3001], MSE_loss:0.09892 epoch [2055/3001], MSE_loss:0.06606 epoch [2056/3001], MSE_loss:0.09516 epoch [2057/3001], MSE_loss:0.11559 epoch [2058/3001], MSE_loss:0.10427 epoch [2059/3001], MSE_loss:0.08768 epoch [2060/3001], MSE_loss:0.06935 epoch [2061/3001], MSE_loss:0.11070 epoch [2062/3001], MSE_loss:0.08024 epoch [2063/3001], MSE_loss:0.08871 epoch [2064/3001], MSE_loss:0.12088 epoch [2065/3001], MSE_loss:0.08546 epoch [2066/3001], MSE_loss:0.08402 epoch [2067/3001], MSE_loss:0.09645 epoch [2068/3001], MSE_loss:0.12822 epoch [2069/3001], MSE_loss:0.11404 epoch [2070/3001], MSE_loss:0.06388 epoch [2071/3001], MSE_loss:0.07378 epoch [2072/3001], MSE_loss:0.09849 epoch [2073/3001], MSE_loss:0.07298 epoch [2074/3001], MSE_loss:0.09222 epoch [2075/3001], MSE_loss:0.10673 epoch [2076/3001], MSE_loss:0.06967 epoch [2077/3001], MSE_loss:0.07641 epoch [2078/3001], MSE_loss:0.10342 epoch [2079/3001], MSE_loss:0.08330 epoch [2080/3001], MSE_loss:0.08370 epoch [2081/3001], MSE_loss:0.08255 epoch [2082/3001], MSE_loss:0.10037 epoch [2083/3001], MSE_loss:0.08434 epoch [2084/3001], MSE_loss:0.09862 epoch [2085/3001], MSE_loss:0.11702 epoch [2086/3001], MSE_loss:0.10012 epoch [2087/3001], MSE_loss:0.12700 epoch [2088/3001], MSE_loss:0.08800 epoch [2089/3001], MSE_loss:0.11461 epoch [2090/3001], MSE_loss:0.12676 epoch [2091/3001], MSE_loss:0.10995 epoch [2092/3001], MSE_loss:0.06846 epoch [2093/3001], MSE_loss:0.08858 epoch [2094/3001], MSE_loss:0.12963 epoch [2095/3001], MSE_loss:0.09696 epoch [2096/3001], MSE_loss:0.08483 epoch [2097/3001], MSE_loss:0.08897 epoch [2098/3001], MSE_loss:0.10182 epoch [2099/3001], MSE_loss:0.10230 epoch [2100/3001], MSE_loss:0.10064 epoch [2101/3001], MSE_loss:0.07917 epoch [2102/3001], MSE_loss:0.08637 epoch [2103/3001], MSE_loss:0.11050 epoch [2104/3001], MSE_loss:0.07124 epoch [2105/3001], MSE_loss:0.07025 epoch [2106/3001], MSE_loss:0.13708 epoch [2107/3001], MSE_loss:0.08910 epoch [2108/3001], MSE_loss:0.08703 epoch [2109/3001], MSE_loss:0.14609 epoch [2110/3001], MSE_loss:0.09648 epoch [2111/3001], MSE_loss:0.10567 epoch [2112/3001], MSE_loss:0.09868 epoch [2113/3001], MSE_loss:0.07424 epoch [2114/3001], MSE_loss:0.09475 epoch [2115/3001], MSE_loss:0.08685 epoch [2116/3001], MSE_loss:0.09865 epoch [2117/3001], MSE_loss:0.12812 epoch [2118/3001], MSE_loss:0.08388 epoch [2119/3001], MSE_loss:0.07890 epoch [2120/3001], MSE_loss:0.14549 epoch [2121/3001], MSE_loss:0.07806 epoch [2122/3001], MSE_loss:0.09772 epoch [2123/3001], MSE_loss:0.08478 epoch [2124/3001], MSE_loss:0.09419 epoch [2125/3001], MSE_loss:0.10672 epoch [2126/3001], MSE_loss:0.09555 epoch [2127/3001], MSE_loss:0.10384 epoch [2128/3001], MSE_loss:0.12529 epoch [2129/3001], MSE_loss:0.12170 epoch [2130/3001], MSE_loss:0.11496 epoch [2131/3001], MSE_loss:0.07904 epoch [2132/3001], MSE_loss:0.07845 epoch [2133/3001], MSE_loss:0.09391 epoch [2134/3001], MSE_loss:0.06294 epoch [2135/3001], MSE_loss:0.09109 epoch [2136/3001], MSE_loss:0.10486 epoch [2137/3001], MSE_loss:0.09654 epoch [2138/3001], MSE_loss:0.12225 epoch [2139/3001], MSE_loss:0.07641 epoch [2140/3001], MSE_loss:0.11016 epoch [2141/3001], MSE_loss:0.10339 epoch [2142/3001], MSE_loss:0.14815 epoch [2143/3001], MSE_loss:0.10722 epoch [2144/3001], MSE_loss:0.09328 epoch [2145/3001], MSE_loss:0.08978 epoch [2146/3001], MSE_loss:0.07655 epoch [2147/3001], MSE_loss:0.07030 epoch [2148/3001], MSE_loss:0.07249 epoch [2149/3001], MSE_loss:0.13673 epoch [2150/3001], MSE_loss:0.09784 epoch [2151/3001], MSE_loss:0.08593 epoch [2152/3001], MSE_loss:0.05010 epoch [2153/3001], MSE_loss:0.10718 epoch [2154/3001], MSE_loss:0.09520 epoch [2155/3001], MSE_loss:0.08303 epoch [2156/3001], MSE_loss:0.08729 epoch [2157/3001], MSE_loss:0.13397 epoch [2158/3001], MSE_loss:0.08129 epoch [2159/3001], MSE_loss:0.07739 epoch [2160/3001], MSE_loss:0.10539 epoch [2161/3001], MSE_loss:0.10592 epoch [2162/3001], MSE_loss:0.08284 epoch [2163/3001], MSE_loss:0.10194 epoch [2164/3001], MSE_loss:0.11157 epoch [2165/3001], MSE_loss:0.08340 epoch [2166/3001], MSE_loss:0.08749 epoch [2167/3001], MSE_loss:0.08139 epoch [2168/3001], MSE_loss:0.07356 epoch [2169/3001], MSE_loss:0.10708 epoch [2170/3001], MSE_loss:0.07216 epoch [2171/3001], MSE_loss:0.12023 epoch [2172/3001], MSE_loss:0.09747 epoch [2173/3001], MSE_loss:0.08076 epoch [2174/3001], MSE_loss:0.06062 epoch [2175/3001], MSE_loss:0.11591 epoch [2176/3001], MSE_loss:0.08347 epoch [2177/3001], MSE_loss:0.07976 epoch [2178/3001], MSE_loss:0.07922 epoch [2179/3001], MSE_loss:0.07886 epoch [2180/3001], MSE_loss:0.14154 epoch [2181/3001], MSE_loss:0.09741 epoch [2182/3001], MSE_loss:0.08193 epoch [2183/3001], MSE_loss:0.09633 epoch [2184/3001], MSE_loss:0.10078 epoch [2185/3001], MSE_loss:0.08481 epoch [2186/3001], MSE_loss:0.10612 epoch [2187/3001], MSE_loss:0.08918 epoch [2188/3001], MSE_loss:0.08959 epoch [2189/3001], MSE_loss:0.08843 epoch [2190/3001], MSE_loss:0.10927 epoch [2191/3001], MSE_loss:0.06953 epoch [2192/3001], MSE_loss:0.10179 epoch [2193/3001], MSE_loss:0.08573 epoch [2194/3001], MSE_loss:0.07538 epoch [2195/3001], MSE_loss:0.11027 epoch [2196/3001], MSE_loss:0.08131 epoch [2197/3001], MSE_loss:0.09751 epoch [2198/3001], MSE_loss:0.09539 epoch [2199/3001], MSE_loss:0.12641 epoch [2200/3001], MSE_loss:0.09253 epoch [2201/3001], MSE_loss:0.09843 epoch [2202/3001], MSE_loss:0.08750 epoch [2203/3001], MSE_loss:0.13134 epoch [2204/3001], MSE_loss:0.08066 epoch [2205/3001], MSE_loss:0.08549 epoch [2206/3001], MSE_loss:0.11819 epoch [2207/3001], MSE_loss:0.09876 epoch [2208/3001], MSE_loss:0.13071 epoch [2209/3001], MSE_loss:0.06531 epoch [2210/3001], MSE_loss:0.09483 epoch [2211/3001], MSE_loss:0.05992 epoch [2212/3001], MSE_loss:0.10085 epoch [2213/3001], MSE_loss:0.08079 epoch [2214/3001], MSE_loss:0.08249 epoch [2215/3001], MSE_loss:0.08226 epoch [2216/3001], MSE_loss:0.08756 epoch [2217/3001], MSE_loss:0.06236 epoch [2218/3001], MSE_loss:0.10009 epoch [2219/3001], MSE_loss:0.08470 epoch [2220/3001], MSE_loss:0.10365 epoch [2221/3001], MSE_loss:0.09959 epoch [2222/3001], MSE_loss:0.10386 epoch [2223/3001], MSE_loss:0.08313 epoch [2224/3001], MSE_loss:0.09691 epoch [2225/3001], MSE_loss:0.13560 epoch [2226/3001], MSE_loss:0.10015 epoch [2227/3001], MSE_loss:0.12552 epoch [2228/3001], MSE_loss:0.09974 epoch [2229/3001], MSE_loss:0.08569 epoch [2230/3001], MSE_loss:0.16534 epoch [2231/3001], MSE_loss:0.07805 epoch [2232/3001], MSE_loss:0.08393 epoch [2233/3001], MSE_loss:0.06660 epoch [2234/3001], MSE_loss:0.09521 epoch [2235/3001], MSE_loss:0.10867 epoch [2236/3001], MSE_loss:0.12557 epoch [2237/3001], MSE_loss:0.09420 epoch [2238/3001], MSE_loss:0.10708 epoch [2239/3001], MSE_loss:0.09299 epoch [2240/3001], MSE_loss:0.06806 epoch [2241/3001], MSE_loss:0.10984 epoch [2242/3001], MSE_loss:0.08583 epoch [2243/3001], MSE_loss:0.07278 epoch [2244/3001], MSE_loss:0.11038 epoch [2245/3001], MSE_loss:0.11533 epoch [2246/3001], MSE_loss:0.07974 epoch [2247/3001], MSE_loss:0.07584 epoch [2248/3001], MSE_loss:0.11822 epoch [2249/3001], MSE_loss:0.08397 epoch [2250/3001], MSE_loss:0.07417 epoch [2251/3001], MSE_loss:0.10122 epoch [2252/3001], MSE_loss:0.05697 epoch [2253/3001], MSE_loss:0.08485 epoch [2254/3001], MSE_loss:0.08993 epoch [2255/3001], MSE_loss:0.11940 epoch [2256/3001], MSE_loss:0.08020 epoch [2257/3001], MSE_loss:0.12142 epoch [2258/3001], MSE_loss:0.14471 epoch [2259/3001], MSE_loss:0.08060 epoch [2260/3001], MSE_loss:0.08041 epoch [2261/3001], MSE_loss:0.06738 epoch [2262/3001], MSE_loss:0.11227 epoch [2263/3001], MSE_loss:0.08865 epoch [2264/3001], MSE_loss:0.09880 epoch [2265/3001], MSE_loss:0.10429 epoch [2266/3001], MSE_loss:0.10647 epoch [2267/3001], MSE_loss:0.08609 epoch [2268/3001], MSE_loss:0.10402 epoch [2269/3001], MSE_loss:0.10403 epoch [2270/3001], MSE_loss:0.08345 epoch [2271/3001], MSE_loss:0.09555 epoch [2272/3001], MSE_loss:0.10017 epoch [2273/3001], MSE_loss:0.14229 epoch [2274/3001], MSE_loss:0.09554 epoch [2275/3001], MSE_loss:0.08804 epoch [2276/3001], MSE_loss:0.09047 epoch [2277/3001], MSE_loss:0.08958 epoch [2278/3001], MSE_loss:0.11801 epoch [2279/3001], MSE_loss:0.10132 epoch [2280/3001], MSE_loss:0.09070 epoch [2281/3001], MSE_loss:0.07434 epoch [2282/3001], MSE_loss:0.07983 epoch [2283/3001], MSE_loss:0.08216 epoch [2284/3001], MSE_loss:0.11883 epoch [2285/3001], MSE_loss:0.09912 epoch [2286/3001], MSE_loss:0.11785 epoch [2287/3001], MSE_loss:0.11201 epoch [2288/3001], MSE_loss:0.12207 epoch [2289/3001], MSE_loss:0.10492 epoch [2290/3001], MSE_loss:0.08641 epoch [2291/3001], MSE_loss:0.13951 epoch [2292/3001], MSE_loss:0.08316 epoch [2293/3001], MSE_loss:0.10520 epoch [2294/3001], MSE_loss:0.12237 epoch [2295/3001], MSE_loss:0.08386 epoch [2296/3001], MSE_loss:0.09166 epoch [2297/3001], MSE_loss:0.09200 epoch [2298/3001], MSE_loss:0.09260 epoch [2299/3001], MSE_loss:0.07025 epoch [2300/3001], MSE_loss:0.08275 epoch [2301/3001], MSE_loss:0.13405 epoch [2302/3001], MSE_loss:0.07617 epoch [2303/3001], MSE_loss:0.14084 epoch [2304/3001], MSE_loss:0.09987 epoch [2305/3001], MSE_loss:0.09929 epoch [2306/3001], MSE_loss:0.09359 epoch [2307/3001], MSE_loss:0.07606 epoch [2308/3001], MSE_loss:0.09855 epoch [2309/3001], MSE_loss:0.08949 epoch [2310/3001], MSE_loss:0.10344 epoch [2311/3001], MSE_loss:0.08799 epoch [2312/3001], MSE_loss:0.08267 epoch [2313/3001], MSE_loss:0.09522 epoch [2314/3001], MSE_loss:0.12368 epoch [2315/3001], MSE_loss:0.11687 epoch [2316/3001], MSE_loss:0.09972 epoch [2317/3001], MSE_loss:0.09950 epoch [2318/3001], MSE_loss:0.09625 epoch [2319/3001], MSE_loss:0.13916 epoch [2320/3001], MSE_loss:0.07325 epoch [2321/3001], MSE_loss:0.10193 epoch [2322/3001], MSE_loss:0.13198 epoch [2323/3001], MSE_loss:0.12256 epoch [2324/3001], MSE_loss:0.08537 epoch [2325/3001], MSE_loss:0.08967 epoch [2326/3001], MSE_loss:0.08611 epoch [2327/3001], MSE_loss:0.08379 epoch [2328/3001], MSE_loss:0.06387 epoch [2329/3001], MSE_loss:0.06848 epoch [2330/3001], MSE_loss:0.06830 epoch [2331/3001], MSE_loss:0.11329 epoch [2332/3001], MSE_loss:0.09928 epoch [2333/3001], MSE_loss:0.09570 epoch [2334/3001], MSE_loss:0.08512 epoch [2335/3001], MSE_loss:0.10270 epoch [2336/3001], MSE_loss:0.07943 epoch [2337/3001], MSE_loss:0.07450 epoch [2338/3001], MSE_loss:0.06189 epoch [2339/3001], MSE_loss:0.08347 epoch [2340/3001], MSE_loss:0.08789 epoch [2341/3001], MSE_loss:0.10049 epoch [2342/3001], MSE_loss:0.08100 epoch [2343/3001], MSE_loss:0.11837 epoch [2344/3001], MSE_loss:0.11187 epoch [2345/3001], MSE_loss:0.07584 epoch [2346/3001], MSE_loss:0.10090 epoch [2347/3001], MSE_loss:0.09580 epoch [2348/3001], MSE_loss:0.10403 epoch [2349/3001], MSE_loss:0.07438 epoch [2350/3001], MSE_loss:0.15336 epoch [2351/3001], MSE_loss:0.12629 epoch [2352/3001], MSE_loss:0.08614 epoch [2353/3001], MSE_loss:0.08289 epoch [2354/3001], MSE_loss:0.10012 epoch [2355/3001], MSE_loss:0.08915 epoch [2356/3001], MSE_loss:0.08596 epoch [2357/3001], MSE_loss:0.09429 epoch [2358/3001], MSE_loss:0.08676 epoch [2359/3001], MSE_loss:0.09406 epoch [2360/3001], MSE_loss:0.08955 epoch [2361/3001], MSE_loss:0.10380 epoch [2362/3001], MSE_loss:0.10035 epoch [2363/3001], MSE_loss:0.07960 epoch [2364/3001], MSE_loss:0.07021 epoch [2365/3001], MSE_loss:0.11321 epoch [2366/3001], MSE_loss:0.09845 epoch [2367/3001], MSE_loss:0.08504 epoch [2368/3001], MSE_loss:0.09314 epoch [2369/3001], MSE_loss:0.12176 epoch [2370/3001], MSE_loss:0.13420 epoch [2371/3001], MSE_loss:0.12622 epoch [2372/3001], MSE_loss:0.10616 epoch [2373/3001], MSE_loss:0.11086 epoch [2374/3001], MSE_loss:0.10073 epoch [2375/3001], MSE_loss:0.07972 epoch [2376/3001], MSE_loss:0.08206 epoch [2377/3001], MSE_loss:0.12416 epoch [2378/3001], MSE_loss:0.08803 epoch [2379/3001], MSE_loss:0.09310 epoch [2380/3001], MSE_loss:0.06376 epoch [2381/3001], MSE_loss:0.09121 epoch [2382/3001], MSE_loss:0.08690 epoch [2383/3001], MSE_loss:0.11346 epoch [2384/3001], MSE_loss:0.12249 epoch [2385/3001], MSE_loss:0.08557 epoch [2386/3001], MSE_loss:0.11089 epoch [2387/3001], MSE_loss:0.08133 epoch [2388/3001], MSE_loss:0.10280 epoch [2389/3001], MSE_loss:0.11542 epoch [2390/3001], MSE_loss:0.12408 epoch [2391/3001], MSE_loss:0.06794 epoch [2392/3001], MSE_loss:0.13073 epoch [2393/3001], MSE_loss:0.09209 epoch [2394/3001], MSE_loss:0.08545 epoch [2395/3001], MSE_loss:0.10342 epoch [2396/3001], MSE_loss:0.09902 epoch [2397/3001], MSE_loss:0.11119 epoch [2398/3001], MSE_loss:0.07865 epoch [2399/3001], MSE_loss:0.08201 epoch [2400/3001], MSE_loss:0.08390 epoch [2401/3001], MSE_loss:0.10880 epoch [2402/3001], MSE_loss:0.08881 epoch [2403/3001], MSE_loss:0.06187 epoch [2404/3001], MSE_loss:0.08444 epoch [2405/3001], MSE_loss:0.10221 epoch [2406/3001], MSE_loss:0.09552 epoch [2407/3001], MSE_loss:0.08147 epoch [2408/3001], MSE_loss:0.10987 epoch [2409/3001], MSE_loss:0.08558 epoch [2410/3001], MSE_loss:0.13123 epoch [2411/3001], MSE_loss:0.10675 epoch [2412/3001], MSE_loss:0.10591 epoch [2413/3001], MSE_loss:0.11338 epoch [2414/3001], MSE_loss:0.11828 epoch [2415/3001], MSE_loss:0.11050 epoch [2416/3001], MSE_loss:0.08908 epoch [2417/3001], MSE_loss:0.07836 epoch [2418/3001], MSE_loss:0.08386 epoch [2419/3001], MSE_loss:0.07516 epoch [2420/3001], MSE_loss:0.07938 epoch [2421/3001], MSE_loss:0.10692 epoch [2422/3001], MSE_loss:0.09832 epoch [2423/3001], MSE_loss:0.10226 epoch [2424/3001], MSE_loss:0.11915 epoch [2425/3001], MSE_loss:0.09480 epoch [2426/3001], MSE_loss:0.07821 epoch [2427/3001], MSE_loss:0.06434 epoch [2428/3001], MSE_loss:0.09009 epoch [2429/3001], MSE_loss:0.08625 epoch [2430/3001], MSE_loss:0.10918 epoch [2431/3001], MSE_loss:0.08796 epoch [2432/3001], MSE_loss:0.06827 epoch [2433/3001], MSE_loss:0.07629 epoch [2434/3001], MSE_loss:0.06022 epoch [2435/3001], MSE_loss:0.08456 epoch [2436/3001], MSE_loss:0.09658 epoch [2437/3001], MSE_loss:0.07115 epoch [2438/3001], MSE_loss:0.07524 epoch [2439/3001], MSE_loss:0.07550 epoch [2440/3001], MSE_loss:0.09003 epoch [2441/3001], MSE_loss:0.08515 epoch [2442/3001], MSE_loss:0.10374 epoch [2443/3001], MSE_loss:0.12911 epoch [2444/3001], MSE_loss:0.12473 epoch [2445/3001], MSE_loss:0.13067 epoch [2446/3001], MSE_loss:0.05369 epoch [2447/3001], MSE_loss:0.09628 epoch [2448/3001], MSE_loss:0.08684 epoch [2449/3001], MSE_loss:0.14370 epoch [2450/3001], MSE_loss:0.11494 epoch [2451/3001], MSE_loss:0.12347 epoch [2452/3001], MSE_loss:0.07900 epoch [2453/3001], MSE_loss:0.10338 epoch [2454/3001], MSE_loss:0.04847 epoch [2455/3001], MSE_loss:0.11066 epoch [2456/3001], MSE_loss:0.13017 epoch [2457/3001], MSE_loss:0.09177 epoch [2458/3001], MSE_loss:0.07979 epoch [2459/3001], MSE_loss:0.07564 epoch [2460/3001], MSE_loss:0.07793 epoch [2461/3001], MSE_loss:0.08846 epoch [2462/3001], MSE_loss:0.08115 epoch [2463/3001], MSE_loss:0.09024 epoch [2464/3001], MSE_loss:0.12978 epoch [2465/3001], MSE_loss:0.09217 epoch [2466/3001], MSE_loss:0.09333 epoch [2467/3001], MSE_loss:0.08281 epoch [2468/3001], MSE_loss:0.09030 epoch [2469/3001], MSE_loss:0.08634 epoch [2470/3001], MSE_loss:0.08878 epoch [2471/3001], MSE_loss:0.11096 epoch [2472/3001], MSE_loss:0.10544 epoch [2473/3001], MSE_loss:0.09675 epoch [2474/3001], MSE_loss:0.09691 epoch [2475/3001], MSE_loss:0.08531 epoch [2476/3001], MSE_loss:0.08402 epoch [2477/3001], MSE_loss:0.09337 epoch [2478/3001], MSE_loss:0.08664 epoch [2479/3001], MSE_loss:0.05246 epoch [2480/3001], MSE_loss:0.10719 epoch [2481/3001], MSE_loss:0.10364 epoch [2482/3001], MSE_loss:0.09197 epoch [2483/3001], MSE_loss:0.09910 epoch [2484/3001], MSE_loss:0.08301 epoch [2485/3001], MSE_loss:0.08488 epoch [2486/3001], MSE_loss:0.08871 epoch [2487/3001], MSE_loss:0.11246 epoch [2488/3001], MSE_loss:0.10656 epoch [2489/3001], MSE_loss:0.08614 epoch [2490/3001], MSE_loss:0.09332 epoch [2491/3001], MSE_loss:0.08232 epoch [2492/3001], MSE_loss:0.07735 epoch [2493/3001], MSE_loss:0.11882 epoch [2494/3001], MSE_loss:0.06868 epoch [2495/3001], MSE_loss:0.13914 epoch [2496/3001], MSE_loss:0.06995 epoch [2497/3001], MSE_loss:0.08123 epoch [2498/3001], MSE_loss:0.07620 epoch [2499/3001], MSE_loss:0.11496 epoch [2500/3001], MSE_loss:0.06201 epoch [2501/3001], MSE_loss:0.10279 epoch [2502/3001], MSE_loss:0.09677 epoch [2503/3001], MSE_loss:0.08866 epoch [2504/3001], MSE_loss:0.11507 epoch [2505/3001], MSE_loss:0.11996 epoch [2506/3001], MSE_loss:0.11303 epoch [2507/3001], MSE_loss:0.08077 epoch [2508/3001], MSE_loss:0.06981 epoch [2509/3001], MSE_loss:0.09066 epoch [2510/3001], MSE_loss:0.10045 epoch [2511/3001], MSE_loss:0.11455 epoch [2512/3001], MSE_loss:0.07201 epoch [2513/3001], MSE_loss:0.07039 epoch [2514/3001], MSE_loss:0.11600 epoch [2515/3001], MSE_loss:0.07626 epoch [2516/3001], MSE_loss:0.07812 epoch [2517/3001], MSE_loss:0.04956 epoch [2518/3001], MSE_loss:0.09556 epoch [2519/3001], MSE_loss:0.11153 epoch [2520/3001], MSE_loss:0.06720 epoch [2521/3001], MSE_loss:0.09710 epoch [2522/3001], MSE_loss:0.08951 epoch [2523/3001], MSE_loss:0.07620 epoch [2524/3001], MSE_loss:0.07927 epoch [2525/3001], MSE_loss:0.11501 epoch [2526/3001], MSE_loss:0.13446 epoch [2527/3001], MSE_loss:0.12573 epoch [2528/3001], MSE_loss:0.08431 epoch [2529/3001], MSE_loss:0.07151 epoch [2530/3001], MSE_loss:0.11227 epoch [2531/3001], MSE_loss:0.09717 epoch [2532/3001], MSE_loss:0.07622 epoch [2533/3001], MSE_loss:0.08941 epoch [2534/3001], MSE_loss:0.09528 epoch [2535/3001], MSE_loss:0.08445 epoch [2536/3001], MSE_loss:0.10585 epoch [2537/3001], MSE_loss:0.10601 epoch [2538/3001], MSE_loss:0.07628 epoch [2539/3001], MSE_loss:0.12351 epoch [2540/3001], MSE_loss:0.08431 epoch [2541/3001], MSE_loss:0.11068 epoch [2542/3001], MSE_loss:0.09718 epoch [2543/3001], MSE_loss:0.06419 epoch [2544/3001], MSE_loss:0.14082 epoch [2545/3001], MSE_loss:0.10313 epoch [2546/3001], MSE_loss:0.07104 epoch [2547/3001], MSE_loss:0.07923 epoch [2548/3001], MSE_loss:0.10556 epoch [2549/3001], MSE_loss:0.08938 epoch [2550/3001], MSE_loss:0.08467 epoch [2551/3001], MSE_loss:0.14977 epoch [2552/3001], MSE_loss:0.13019 epoch [2553/3001], MSE_loss:0.11939 epoch [2554/3001], MSE_loss:0.10012 epoch [2555/3001], MSE_loss:0.11082 epoch [2556/3001], MSE_loss:0.09529 epoch [2557/3001], MSE_loss:0.09679 epoch [2558/3001], MSE_loss:0.15171 epoch [2559/3001], MSE_loss:0.07668 epoch [2560/3001], MSE_loss:0.09909 epoch [2561/3001], MSE_loss:0.08601 epoch [2562/3001], MSE_loss:0.08443 epoch [2563/3001], MSE_loss:0.07880 epoch [2564/3001], MSE_loss:0.10200 epoch [2565/3001], MSE_loss:0.12885 epoch [2566/3001], MSE_loss:0.09943 epoch [2567/3001], MSE_loss:0.08842 epoch [2568/3001], MSE_loss:0.08141 epoch [2569/3001], MSE_loss:0.14038 epoch [2570/3001], MSE_loss:0.11453 epoch [2571/3001], MSE_loss:0.08416 epoch [2572/3001], MSE_loss:0.10765 epoch [2573/3001], MSE_loss:0.10828 epoch [2574/3001], MSE_loss:0.11605 epoch [2575/3001], MSE_loss:0.10426 epoch [2576/3001], MSE_loss:0.09796 epoch [2577/3001], MSE_loss:0.14486 epoch [2578/3001], MSE_loss:0.08937 epoch [2579/3001], MSE_loss:0.11220 epoch [2580/3001], MSE_loss:0.08195 epoch [2581/3001], MSE_loss:0.07958 epoch [2582/3001], MSE_loss:0.07873 epoch [2583/3001], MSE_loss:0.12853 epoch [2584/3001], MSE_loss:0.08783 epoch [2585/3001], MSE_loss:0.12084 epoch [2586/3001], MSE_loss:0.11583 epoch [2587/3001], MSE_loss:0.10693 epoch [2588/3001], MSE_loss:0.08593 epoch [2589/3001], MSE_loss:0.08899 epoch [2590/3001], MSE_loss:0.10912 epoch [2591/3001], MSE_loss:0.14982 epoch [2592/3001], MSE_loss:0.07328 epoch [2593/3001], MSE_loss:0.11006 epoch [2594/3001], MSE_loss:0.09528 epoch [2595/3001], MSE_loss:0.09268 epoch [2596/3001], MSE_loss:0.11403 epoch [2597/3001], MSE_loss:0.11904 epoch [2598/3001], MSE_loss:0.08288 epoch [2599/3001], MSE_loss:0.11617 epoch [2600/3001], MSE_loss:0.07685 epoch [2601/3001], MSE_loss:0.10172 epoch [2602/3001], MSE_loss:0.10112 epoch [2603/3001], MSE_loss:0.09042 epoch [2604/3001], MSE_loss:0.09650 epoch [2605/3001], MSE_loss:0.11043 epoch [2606/3001], MSE_loss:0.10036 epoch [2607/3001], MSE_loss:0.11448 epoch [2608/3001], MSE_loss:0.08604 epoch [2609/3001], MSE_loss:0.05653 epoch [2610/3001], MSE_loss:0.12615 epoch [2611/3001], MSE_loss:0.08781 epoch [2612/3001], MSE_loss:0.06844 epoch [2613/3001], MSE_loss:0.09953 epoch [2614/3001], MSE_loss:0.06992 epoch [2615/3001], MSE_loss:0.09575 epoch [2616/3001], MSE_loss:0.09303 epoch [2617/3001], MSE_loss:0.06585 epoch [2618/3001], MSE_loss:0.11832 epoch [2619/3001], MSE_loss:0.06896 epoch [2620/3001], MSE_loss:0.09064 epoch [2621/3001], MSE_loss:0.07913 epoch [2622/3001], MSE_loss:0.08562 epoch [2623/3001], MSE_loss:0.09941 epoch [2624/3001], MSE_loss:0.13539 epoch [2625/3001], MSE_loss:0.10541 epoch [2626/3001], MSE_loss:0.06349 epoch [2627/3001], MSE_loss:0.10066 epoch [2628/3001], MSE_loss:0.13124 epoch [2629/3001], MSE_loss:0.07099 epoch [2630/3001], MSE_loss:0.08960 epoch [2631/3001], MSE_loss:0.10399 epoch [2632/3001], MSE_loss:0.07750 epoch [2633/3001], MSE_loss:0.05440 epoch [2634/3001], MSE_loss:0.11897 epoch [2635/3001], MSE_loss:0.09237 epoch [2636/3001], MSE_loss:0.09045 epoch [2637/3001], MSE_loss:0.10986 epoch [2638/3001], MSE_loss:0.08432 epoch [2639/3001], MSE_loss:0.09434 epoch [2640/3001], MSE_loss:0.05912 epoch [2641/3001], MSE_loss:0.08494 epoch [2642/3001], MSE_loss:0.09026 epoch [2643/3001], MSE_loss:0.09391 epoch [2644/3001], MSE_loss:0.06038 epoch [2645/3001], MSE_loss:0.12921 epoch [2646/3001], MSE_loss:0.06351 epoch [2647/3001], MSE_loss:0.07427 epoch [2648/3001], MSE_loss:0.11495 epoch [2649/3001], MSE_loss:0.12778 epoch [2650/3001], MSE_loss:0.11957 epoch [2651/3001], MSE_loss:0.07450 epoch [2652/3001], MSE_loss:0.08479 epoch [2653/3001], MSE_loss:0.07835 epoch [2654/3001], MSE_loss:0.08390 epoch [2655/3001], MSE_loss:0.06833 epoch [2656/3001], MSE_loss:0.09060 epoch [2657/3001], MSE_loss:0.09810 epoch [2658/3001], MSE_loss:0.08160 epoch [2659/3001], MSE_loss:0.09888 epoch [2660/3001], MSE_loss:0.08925 epoch [2661/3001], MSE_loss:0.07856 epoch [2662/3001], MSE_loss:0.12417 epoch [2663/3001], MSE_loss:0.09452 epoch [2664/3001], MSE_loss:0.10241 epoch [2665/3001], MSE_loss:0.07942 epoch [2666/3001], MSE_loss:0.08471 epoch [2667/3001], MSE_loss:0.07629 epoch [2668/3001], MSE_loss:0.08796 epoch [2669/3001], MSE_loss:0.08519 epoch [2670/3001], MSE_loss:0.09536 epoch [2671/3001], MSE_loss:0.07444 epoch [2672/3001], MSE_loss:0.10359 epoch [2673/3001], MSE_loss:0.07683 epoch [2674/3001], MSE_loss:0.12450 epoch [2675/3001], MSE_loss:0.08346 epoch [2676/3001], MSE_loss:0.10259 epoch [2677/3001], MSE_loss:0.10770 epoch [2678/3001], MSE_loss:0.07255 epoch [2679/3001], MSE_loss:0.09694 epoch [2680/3001], MSE_loss:0.10852 epoch [2681/3001], MSE_loss:0.14563 epoch [2682/3001], MSE_loss:0.12756 epoch [2683/3001], MSE_loss:0.10185 epoch [2684/3001], MSE_loss:0.11616 epoch [2685/3001], MSE_loss:0.08004 epoch [2686/3001], MSE_loss:0.11460 epoch [2687/3001], MSE_loss:0.14258 epoch [2688/3001], MSE_loss:0.08259 epoch [2689/3001], MSE_loss:0.15992 epoch [2690/3001], MSE_loss:0.09696 epoch [2691/3001], MSE_loss:0.07530 epoch [2692/3001], MSE_loss:0.10202 epoch [2693/3001], MSE_loss:0.11844 epoch [2694/3001], MSE_loss:0.10907 epoch [2695/3001], MSE_loss:0.13215 epoch [2696/3001], MSE_loss:0.07921 epoch [2697/3001], MSE_loss:0.10788 epoch [2698/3001], MSE_loss:0.06571 epoch [2699/3001], MSE_loss:0.11228 epoch [2700/3001], MSE_loss:0.09816 epoch [2701/3001], MSE_loss:0.07927 epoch [2702/3001], MSE_loss:0.06994 epoch [2703/3001], MSE_loss:0.06109 epoch [2704/3001], MSE_loss:0.11398 epoch [2705/3001], MSE_loss:0.09276 epoch [2706/3001], MSE_loss:0.06773 epoch [2707/3001], MSE_loss:0.08761 epoch [2708/3001], MSE_loss:0.10046 epoch [2709/3001], MSE_loss:0.07126 epoch [2710/3001], MSE_loss:0.10551 epoch [2711/3001], MSE_loss:0.11853 epoch [2712/3001], MSE_loss:0.05703 epoch [2713/3001], MSE_loss:0.07015 epoch [2714/3001], MSE_loss:0.09212 epoch [2715/3001], MSE_loss:0.11270 epoch [2716/3001], MSE_loss:0.12030 epoch [2717/3001], MSE_loss:0.11233 epoch [2718/3001], MSE_loss:0.06922 epoch [2719/3001], MSE_loss:0.08930 epoch [2720/3001], MSE_loss:0.12352 epoch [2721/3001], MSE_loss:0.09298 epoch [2722/3001], MSE_loss:0.07133 epoch [2723/3001], MSE_loss:0.08186 epoch [2724/3001], MSE_loss:0.08194 epoch [2725/3001], MSE_loss:0.06921 epoch [2726/3001], MSE_loss:0.07788 epoch [2727/3001], MSE_loss:0.08638 epoch [2728/3001], MSE_loss:0.09983 epoch [2729/3001], MSE_loss:0.11769 epoch [2730/3001], MSE_loss:0.07953 epoch [2731/3001], MSE_loss:0.09570 epoch [2732/3001], MSE_loss:0.12427 epoch [2733/3001], MSE_loss:0.16173 epoch [2734/3001], MSE_loss:0.07395 epoch [2735/3001], MSE_loss:0.08274 epoch [2736/3001], MSE_loss:0.11343 epoch [2737/3001], MSE_loss:0.08002 epoch [2738/3001], MSE_loss:0.06826 epoch [2739/3001], MSE_loss:0.11699 epoch [2740/3001], MSE_loss:0.07041 epoch [2741/3001], MSE_loss:0.11093 epoch [2742/3001], MSE_loss:0.09311 epoch [2743/3001], MSE_loss:0.10739 epoch [2744/3001], MSE_loss:0.08028 epoch [2745/3001], MSE_loss:0.10044 epoch [2746/3001], MSE_loss:0.08103 epoch [2747/3001], MSE_loss:0.06918 epoch [2748/3001], MSE_loss:0.08589 epoch [2749/3001], MSE_loss:0.08971 epoch [2750/3001], MSE_loss:0.07638 epoch [2751/3001], MSE_loss:0.10961 epoch [2752/3001], MSE_loss:0.10781 epoch [2753/3001], MSE_loss:0.09235 epoch [2754/3001], MSE_loss:0.09534 epoch [2755/3001], MSE_loss:0.10644 epoch [2756/3001], MSE_loss:0.11037 epoch [2757/3001], MSE_loss:0.07514 epoch [2758/3001], MSE_loss:0.15115 epoch [2759/3001], MSE_loss:0.08964 epoch [2760/3001], MSE_loss:0.15431 epoch [2761/3001], MSE_loss:0.09682 epoch [2762/3001], MSE_loss:0.08562 epoch [2763/3001], MSE_loss:0.10393 epoch [2764/3001], MSE_loss:0.08962 epoch [2765/3001], MSE_loss:0.06566 epoch [2766/3001], MSE_loss:0.10636 epoch [2767/3001], MSE_loss:0.12883 epoch [2768/3001], MSE_loss:0.09941 epoch [2769/3001], MSE_loss:0.07585 epoch [2770/3001], MSE_loss:0.06529 epoch [2771/3001], MSE_loss:0.08892 epoch [2772/3001], MSE_loss:0.13775 epoch [2773/3001], MSE_loss:0.09152 epoch [2774/3001], MSE_loss:0.09305 epoch [2775/3001], MSE_loss:0.14591 epoch [2776/3001], MSE_loss:0.09781 epoch [2777/3001], MSE_loss:0.09979 epoch [2778/3001], MSE_loss:0.13632 epoch [2779/3001], MSE_loss:0.07362 epoch [2780/3001], MSE_loss:0.05946 epoch [2781/3001], MSE_loss:0.05945 epoch [2782/3001], MSE_loss:0.12903 epoch [2783/3001], MSE_loss:0.07703 epoch [2784/3001], MSE_loss:0.11276 epoch [2785/3001], MSE_loss:0.10525 epoch [2786/3001], MSE_loss:0.09045 epoch [2787/3001], MSE_loss:0.13020 epoch [2788/3001], MSE_loss:0.10076 epoch [2789/3001], MSE_loss:0.09241 epoch [2790/3001], MSE_loss:0.09677 epoch [2791/3001], MSE_loss:0.07949 epoch [2792/3001], MSE_loss:0.11726 epoch [2793/3001], MSE_loss:0.11945 epoch [2794/3001], MSE_loss:0.06456 epoch [2795/3001], MSE_loss:0.09451 epoch [2796/3001], MSE_loss:0.08303 epoch [2797/3001], MSE_loss:0.10154 epoch [2798/3001], MSE_loss:0.10597 epoch [2799/3001], MSE_loss:0.09102 epoch [2800/3001], MSE_loss:0.09809 epoch [2801/3001], MSE_loss:0.11943 epoch [2802/3001], MSE_loss:0.10561 epoch [2803/3001], MSE_loss:0.09108 epoch [2804/3001], MSE_loss:0.14619 epoch [2805/3001], MSE_loss:0.04291 epoch [2806/3001], MSE_loss:0.12017 epoch [2807/3001], MSE_loss:0.09988 epoch [2808/3001], MSE_loss:0.09331 epoch [2809/3001], MSE_loss:0.07819 epoch [2810/3001], MSE_loss:0.08890 epoch [2811/3001], MSE_loss:0.08555 epoch [2812/3001], MSE_loss:0.10897 epoch [2813/3001], MSE_loss:0.08070 epoch [2814/3001], MSE_loss:0.06970 epoch [2815/3001], MSE_loss:0.08584 epoch [2816/3001], MSE_loss:0.10541 epoch [2817/3001], MSE_loss:0.08027 epoch [2818/3001], MSE_loss:0.10084 epoch [2819/3001], MSE_loss:0.08973 epoch [2820/3001], MSE_loss:0.08614 epoch [2821/3001], MSE_loss:0.09167 epoch [2822/3001], MSE_loss:0.11799 epoch [2823/3001], MSE_loss:0.09297 epoch [2824/3001], MSE_loss:0.15218 epoch [2825/3001], MSE_loss:0.09988 epoch [2826/3001], MSE_loss:0.10393 epoch [2827/3001], MSE_loss:0.10248 epoch [2828/3001], MSE_loss:0.09625 epoch [2829/3001], MSE_loss:0.09872 epoch [2830/3001], MSE_loss:0.08173 epoch [2831/3001], MSE_loss:0.09654 epoch [2832/3001], MSE_loss:0.08024 epoch [2833/3001], MSE_loss:0.09534 epoch [2834/3001], MSE_loss:0.06978 epoch [2835/3001], MSE_loss:0.07472 epoch [2836/3001], MSE_loss:0.11010 epoch [2837/3001], MSE_loss:0.10670 epoch [2838/3001], MSE_loss:0.10510 epoch [2839/3001], MSE_loss:0.07297 epoch [2840/3001], MSE_loss:0.10450 epoch [2841/3001], MSE_loss:0.07452 epoch [2842/3001], MSE_loss:0.09950 epoch [2843/3001], MSE_loss:0.11525 epoch [2844/3001], MSE_loss:0.13632 epoch [2845/3001], MSE_loss:0.07012 epoch [2846/3001], MSE_loss:0.05496 epoch [2847/3001], MSE_loss:0.10573 epoch [2848/3001], MSE_loss:0.08814 epoch [2849/3001], MSE_loss:0.10005 epoch [2850/3001], MSE_loss:0.10705 epoch [2851/3001], MSE_loss:0.13118 epoch [2852/3001], MSE_loss:0.12874 epoch [2853/3001], MSE_loss:0.08448 epoch [2854/3001], MSE_loss:0.07058 epoch [2855/3001], MSE_loss:0.12087 epoch [2856/3001], MSE_loss:0.13575 epoch [2857/3001], MSE_loss:0.08098 epoch [2858/3001], MSE_loss:0.08666 epoch [2859/3001], MSE_loss:0.10862 epoch [2860/3001], MSE_loss:0.06903 epoch [2861/3001], MSE_loss:0.11940 epoch [2862/3001], MSE_loss:0.10472 epoch [2863/3001], MSE_loss:0.09343 epoch [2864/3001], MSE_loss:0.06725 epoch [2865/3001], MSE_loss:0.09528 epoch [2866/3001], MSE_loss:0.08233 epoch [2867/3001], MSE_loss:0.06243 epoch [2868/3001], MSE_loss:0.08371 epoch [2869/3001], MSE_loss:0.11563 epoch [2870/3001], MSE_loss:0.09839 epoch [2871/3001], MSE_loss:0.08725 epoch [2872/3001], MSE_loss:0.10562 epoch [2873/3001], MSE_loss:0.14787 epoch [2874/3001], MSE_loss:0.08297 epoch [2875/3001], MSE_loss:0.12115 epoch [2876/3001], MSE_loss:0.10420 epoch [2877/3001], MSE_loss:0.09896 epoch [2878/3001], MSE_loss:0.06696 epoch [2879/3001], MSE_loss:0.06017 epoch [2880/3001], MSE_loss:0.12381 epoch [2881/3001], MSE_loss:0.06842 epoch [2882/3001], MSE_loss:0.07849 epoch [2883/3001], MSE_loss:0.10305 epoch [2884/3001], MSE_loss:0.09782 epoch [2885/3001], MSE_loss:0.10213 epoch [2886/3001], MSE_loss:0.13181 epoch [2887/3001], MSE_loss:0.11646 epoch [2888/3001], MSE_loss:0.09134 epoch [2889/3001], MSE_loss:0.10866 epoch [2890/3001], MSE_loss:0.07108 epoch [2891/3001], MSE_loss:0.07661 epoch [2892/3001], MSE_loss:0.11409 epoch [2893/3001], MSE_loss:0.10808 epoch [2894/3001], MSE_loss:0.11729 epoch [2895/3001], MSE_loss:0.10530 epoch [2896/3001], MSE_loss:0.08119 epoch [2897/3001], MSE_loss:0.10320 epoch [2898/3001], MSE_loss:0.10105 epoch [2899/3001], MSE_loss:0.08975 epoch [2900/3001], MSE_loss:0.09367 epoch [2901/3001], MSE_loss:0.08126 epoch [2902/3001], MSE_loss:0.13281 epoch [2903/3001], MSE_loss:0.09927 epoch [2904/3001], MSE_loss:0.09462 epoch [2905/3001], MSE_loss:0.07139 epoch [2906/3001], MSE_loss:0.11473 epoch [2907/3001], MSE_loss:0.08873 epoch [2908/3001], MSE_loss:0.11020 epoch [2909/3001], MSE_loss:0.09354 epoch [2910/3001], MSE_loss:0.08468 epoch [2911/3001], MSE_loss:0.10356 epoch [2912/3001], MSE_loss:0.10884 epoch [2913/3001], MSE_loss:0.11462 epoch [2914/3001], MSE_loss:0.18830 epoch [2915/3001], MSE_loss:0.07741 epoch [2916/3001], MSE_loss:0.12567 epoch [2917/3001], MSE_loss:0.12150 epoch [2918/3001], MSE_loss:0.09022 epoch [2919/3001], MSE_loss:0.10649 epoch [2920/3001], MSE_loss:0.17785 epoch [2921/3001], MSE_loss:0.08234 epoch [2922/3001], MSE_loss:0.07845 epoch [2923/3001], MSE_loss:0.09254 epoch [2924/3001], MSE_loss:0.08318 epoch [2925/3001], MSE_loss:0.13312 epoch [2926/3001], MSE_loss:0.08776 epoch [2927/3001], MSE_loss:0.11536 epoch [2928/3001], MSE_loss:0.08675 epoch [2929/3001], MSE_loss:0.07847 epoch [2930/3001], MSE_loss:0.09596 epoch [2931/3001], MSE_loss:0.09700 epoch [2932/3001], MSE_loss:0.08447 epoch [2933/3001], MSE_loss:0.08735 epoch [2934/3001], MSE_loss:0.09616 epoch [2935/3001], MSE_loss:0.10218 epoch [2936/3001], MSE_loss:0.09042 epoch [2937/3001], MSE_loss:0.10967 epoch [2938/3001], MSE_loss:0.10528 epoch [2939/3001], MSE_loss:0.14068 epoch [2940/3001], MSE_loss:0.11608 epoch [2941/3001], MSE_loss:0.06963 epoch [2942/3001], MSE_loss:0.07776 epoch [2943/3001], MSE_loss:0.11924 epoch [2944/3001], MSE_loss:0.10236 epoch [2945/3001], MSE_loss:0.07521 epoch [2946/3001], MSE_loss:0.09391 epoch [2947/3001], MSE_loss:0.07896 epoch [2948/3001], MSE_loss:0.10231 epoch [2949/3001], MSE_loss:0.09756 epoch [2950/3001], MSE_loss:0.10019 epoch [2951/3001], MSE_loss:0.08406 epoch [2952/3001], MSE_loss:0.10172 epoch [2953/3001], MSE_loss:0.11695 epoch [2954/3001], MSE_loss:0.08091 epoch [2955/3001], MSE_loss:0.10450 epoch [2956/3001], MSE_loss:0.10661 epoch [2957/3001], MSE_loss:0.10810 epoch [2958/3001], MSE_loss:0.06797 epoch [2959/3001], MSE_loss:0.07782 epoch [2960/3001], MSE_loss:0.10804 epoch [2961/3001], MSE_loss:0.08966 epoch [2962/3001], MSE_loss:0.10016 epoch [2963/3001], MSE_loss:0.10854 epoch [2964/3001], MSE_loss:0.10394 epoch [2965/3001], MSE_loss:0.10314 epoch [2966/3001], MSE_loss:0.06847 epoch [2967/3001], MSE_loss:0.09919 epoch [2968/3001], MSE_loss:0.07746 epoch [2969/3001], MSE_loss:0.10994 epoch [2970/3001], MSE_loss:0.08296 epoch [2971/3001], MSE_loss:0.10513 epoch [2972/3001], MSE_loss:0.13482 epoch [2973/3001], MSE_loss:0.11122 epoch [2974/3001], MSE_loss:0.06961 epoch [2975/3001], MSE_loss:0.14837 epoch [2976/3001], MSE_loss:0.09922 epoch [2977/3001], MSE_loss:0.14484 epoch [2978/3001], MSE_loss:0.12465 epoch [2979/3001], MSE_loss:0.10646 epoch [2980/3001], MSE_loss:0.09371 epoch [2981/3001], MSE_loss:0.09852 epoch [2982/3001], MSE_loss:0.06471 epoch [2983/3001], MSE_loss:0.11949 epoch [2984/3001], MSE_loss:0.05555 epoch [2985/3001], MSE_loss:0.11071 epoch [2986/3001], MSE_loss:0.07408 epoch [2987/3001], MSE_loss:0.08415 epoch [2988/3001], MSE_loss:0.11859 epoch [2989/3001], MSE_loss:0.12427 epoch [2990/3001], MSE_loss:0.09785 epoch [2991/3001], MSE_loss:0.12652 epoch [2992/3001], MSE_loss:0.10473 epoch [2993/3001], MSE_loss:0.08018 epoch [2994/3001], MSE_loss:0.10373 epoch [2995/3001], MSE_loss:0.08667 epoch [2996/3001], MSE_loss:0.11422 epoch [2997/3001], MSE_loss:0.11538 epoch [2998/3001], MSE_loss:0.05891 epoch [2999/3001], MSE_loss:0.07824 epoch [3000/3001], MSE_loss:0.08720 epoch [3001/3001], MSE_loss:0.06168
latent features visualization¶
In [ ]:
input_data_device = input_data.to(device)
latent_features = autoencoder.encode(input_data_device).detach().cpu().numpy()
print(latent_features.shape)
(360, 16)
In [ ]:
tsne = TSNE(n_components=2, init='pca', random_state=0, n_jobs=-1, perplexity=30)
tsne_2d = tsne.fit_transform(latent_features)
fig, axi1=plt.subplots(1, figsize=(2, 1.5))
axi1.scatter(tsne_2d[:, 0], tsne_2d[:, 1],
marker='*', s=10, color=sns.color_palette('Paired')[1],
)
ax = plt.gca()
plt.grid(True, linewidth=0.5, color='gray', linestyle=':')
plt.xlim([-30, 30])
plt.ylim([-30, 30])
# plt.yticks([-100,-50,0,50,100],[-100,-50,0,50,100])
# plt.xticks([-100,-50,0,50,100],[-100,-50,0,50,100])
ax.tick_params(which='both', bottom=True, top=False, left=True, right=False,
labelbottom=True, labelleft=True, direction='out',width=1)
plt.show()
Train clustering network¶
In [ ]:
tmp = copy.deepcopy(input_data_raw)
np.random.shuffle(tmp)
np.random.shuffle(tmp)
data_shuffle = torch.from_numpy(tmp).type(torch.FloatTensor)
In [ ]:
EPOCHS = 10001
BATCH_SIZE = 512
lr = 0.6
file_path_prefix = './network_data/'
for ind, n_c in enumerate([6, 8, 10]): #
autoencoder = UDEC_Network.AutoEncoder().to(device)
ae_save_path = file_path_prefix + 'autoencoder.pth'
checkpoint = torch.load(ae_save_path)
autoencoder.load_state_dict(checkpoint['state_dict'])
dec = UDEC_Network.DEC(n_clusters=n_c, autoencoder=autoencoder, hidden=16, cluster_centers=None, alpha=1.0).to(device)
dec_save_path = file_path_prefix + 'dec-' + str(n_c) + '-clusters' + '.pth'
checkpoint = { "epoch": 0, "best": float("inf") }
UDEC_Network.train(data=data_shuffle, model=dec, num_epochs=EPOCHS,
n_cluster=n_c, draw_pic=True, lr=lr, file_path_prefix=file_path_prefix,
savepath=dec_save_path, checkpoint=checkpoint, batch_size=BATCH_SIZE)
Training plotting Epochs: [0/10001] Loss:0.197330504655838 Epochs: [1000/10001] Loss:0.011421783827245235 Epochs: [2000/10001] Loss:0.00821361131966114 Epochs: [3000/10001] Loss:0.006751133594661951 Epochs: [4000/10001] Loss:0.005868184845894575 plotting Epochs: [5000/10001] Loss:0.005261091515421867 Epochs: [6000/10001] Loss:0.004810708109289408 Epochs: [7000/10001] Loss:0.004459407180547714 Epochs: [8000/10001] Loss:0.004175422713160515 Epochs: [9000/10001] Loss:0.003939704969525337 plotting Epochs: [10000/10001] Loss:0.0037399560678750277 Training plotting Epochs: [0/10001] Loss:0.23771421611309052 Epochs: [1000/10001] Loss:0.016643472015857697 Epochs: [2000/10001] Loss:0.011941717006266117 Epochs: [3000/10001] Loss:0.009803962893784046 Epochs: [4000/10001] Loss:0.00851544737815857 plotting Epochs: [5000/10001] Loss:0.007630394771695137 Epochs: [6000/10001] Loss:0.00697431992739439 Epochs: [7000/10001] Loss:0.006462802179157734 Epochs: [8000/10001] Loss:0.006049491930752993 Epochs: [9000/10001] Loss:0.005706531461328268 plotting Epochs: [10000/10001] Loss:0.005415949039161205 Training plotting Epochs: [0/10001] Loss:0.27588802576065063 Epochs: [1000/10001] Loss:0.021534908562898636 Epochs: [2000/10001] Loss:0.015373367816209793 Epochs: [3000/10001] Loss:0.01258988119661808 Epochs: [4000/10001] Loss:0.010917743667960167 plotting Epochs: [5000/10001] Loss:0.00977224763482809 Epochs: [6000/10001] Loss:0.008924301713705063 Epochs: [7000/10001] Loss:0.008263779804110527 Epochs: [8000/10001] Loss:0.0077304840087890625 Epochs: [9000/10001] Loss:0.0072882771492004395 plotting Epochs: [10000/10001] Loss:0.006913915276527405
K-Means clustering of latent features¶
In [ ]:
data = torch.from_numpy(input_data_raw).type(torch.FloatTensor)
num_clusters = [6, 8, 10]
latent_vec = np.empty( (len(num_clusters), data.size()[0], 16) ) # [n_c, frames, n_hidden]
pred_label = np.empty( (len(num_clusters), data.size()[0]) )
pred_center = []
file_path_prefix = './network_data/'
for ind, n_c in enumerate(num_clusters):
# load model
autoencoder = UDEC_Network.AutoEncoder()
dec = UDEC_Network.DEC(n_clusters=n_c, autoencoder=autoencoder, hidden=16, cluster_centers=None, alpha=1.0)
dec_save_path = file_path_prefix + 'dec-' + str(n_c) + '-clusters' + '.pth'
checkpoint = torch.load(dec_save_path)
dec.load_state_dict(checkpoint['state_dict']);
# calculate latent vectors
latent_vec[ind] = dec.autoencoder.encode(input_data).detach().cpu().numpy()
# get cluster centers
cluster = KMeans(n_clusters=num_clusters[ind], random_state=0, n_init=10*num_clusters[ind]).fit(latent_vec[ind])
# get cluster labels
centroid = cluster.cluster_centers_
pred_center.append(centroid)
pred_label[ind] = cluster.labels_
pred_label += 1
In [ ]:
# with open('./tmp_data/eeg_data_udec_clustering_results.pkl', 'wb') as f:
# dill.dump([pred_center, pred_label, centroid, latent_vec], f)
In [ ]:
with open('./tmp_data/eeg_data_udec_clustering_results.pkl', 'rb') as f:
[pred_center, pred_label, centroid, latent_vec] = dill.load(f)
T-SNE visualization¶
In [ ]:
tsne = TSNE(n_components=2, init='pca', random_state=0, n_jobs=-1, perplexity=30)
colors = sns.color_palette('pastel')
for ind, n_c in enumerate(num_clusters):
lf_with_center = np.vstack((latent_vec[ind], pred_center[ind]))
kmeans_2d_with_center = tsne.fit_transform(lf_with_center)
kmeans_2d = kmeans_2d_with_center[:-n_c, :]
kmeans_2d_center = kmeans_2d_with_center[-n_c:, :]
fig, axi1=plt.subplots(1, figsize=(2, 1.5))
for i in range(n_c):
axi1.scatter( kmeans_2d[pred_label[ind] == i+1, 0],
kmeans_2d[pred_label[ind] == i+1, 1],
marker='*',
s=10,
color=plt.cm.tab20(i%20),
)
axi1.text(kmeans_2d_center[i,0], kmeans_2d_center[i,1],
i+1, fontsize=9, fontweight='semibold',
verticalalignment='center', horizontalalignment='center',
color='black',
)
ax = plt.gca()
plt.grid(True, linewidth=0.5, color='gray', linestyle=':')
plt.xlim([-40, 40])
plt.ylim([-40, 40])
# plt.yticks([-100,-50,0,50,100],[-100,-50,0,50,100])
# plt.xticks([-100,-50,0,50,100],[-100,-50,0,50,100])
ax.tick_params(which='both', bottom=True, top=False, left=True, right=False,
labelbottom=True, labelleft=True, direction='out',width=1)
plt.show()
Find index of cluster centers¶
In [ ]:
euc_center_index = []
rie_center_index = []
num_clusters = [6, 8, 10]
riemann_dist_symmetry = np.zeros_like(cov_dist_s)
for i in range(cov_dist_s.shape[0]):
for j in range(i+1, cov_dist_s.shape[0]):
riemann_dist_symmetry[i,j] = cov_dist_s[i,j]
riemann_dist_symmetry[j,i] = cov_dist_s[i,j]
# Euclidean distance of lantent vectors
for i, n_c in enumerate(num_clusters):
tmp_c = []
for n in range(n_c):
distance = np.sum(np.square(latent_vec[i][pred_label[i]==(n+1)] - pred_center[i][n]), axis=1)
index = distance.argmin()
tmp_c.append( np.argwhere(pred_label[i]==(n+1))[index][0] )
euc_center_index.append(tmp_c)
# Riemann distance
for i, n_c in enumerate(num_clusters):
tmp_c = []
for n in range(n_c):
tmp = riemann_dist_symmetry[pred_label[i]==(n+1), :]
tmp2 = tmp[:, pred_label[i]==(n+1)]
distance = np.sum(tmp2, axis=0)
index = distance.argmin()
tmp_c.append( np.argwhere(pred_label[i]==(n+1))[index][0] )
rie_center_index.append(tmp_c)
In [ ]:
print(euc_center_index[1])
print(rie_center_index[1])
[7, 260, 160, 348, 224, 208, 81, 131] [12, 272, 151, 341, 229, 28, 81, 120]
In [ ]:
clu_index = 2 # 10 clusters
cog_condition = [] # 2 conditions
time_index = [] # 0-180 time window
for ci in rie_center_index[clu_index]:
cog_condition.append( 1+int(ci/180) )
time_index.append( ci%180 )
In [ ]:
print(cog_condition)
print(time_index)
[1, 2, 2, 1, 2, 1, 1, 1, 2, 1] [16, 92, 53, 173, 161, 31, 81, 148, 122, 120]
t-test of clustering results¶
In [ ]:
for n_c in range(len(num_clusters)):
p_res = ttest_for_clusters.distance_ttest(pred_label[n_c], cov_dist_s, samp_num=500)
ttest_for_clusters.draw_test_mat(p_res, corr_p=True)
[]
[]
[]
Plot microstate series¶
In [ ]:
cluster_number_index = 2 # 10 clusters
each_condition_label = np.full((2, 180), np.nan)
type_num = 2
for tp in range(type_num):
st = tp*180
ed = st+180
each_condition_label[tp] = pred_label[cluster_number_index, st:ed]
In [ ]:
for sym in range(type_num):
draw_states.draw_state_blocks_for_eeg(each_condition_label[sym], figsize=(3.5, 1.0), n_clusters=10,
# n_clusters=int(np.max(each_condition_label[sym])),
colorbar_fraction=0.015, tmin=0.0, tmax=0.7,
colorbar_ticks=[1,4,7,10]);
In [ ]:
# importlib.reload(draw_states)
for sym in range(type_num):
draw_states.draw_state_blocks_for_eeg(each_condition_label[sym], figsize=(3.5, 1.0), n_clusters=10,
# n_clusters=int(np.max(each_condition_label[sym])),
colorbar_fraction=0.015, tmin=0.0, tmax=0.7,
colorbar_ticks=[1,4,7,10], rie_dist=cov_dist_s, current_cluster=sym);
Draw center EEG topomap¶
In [ ]:
cog_condition = [1, 2, 2, 1, 2, 1, 1, 1, 2, 1]
time_index = [16, 92, 53, 173, 161, 31, 81, 148, 122, 120]
In [ ]:
# erp_data_mean [2, 28, 256]
half_win = 12
start_t = 51
end_t = 231
len_t = 180
eeg_state_centers = np.zeros((10, 28))
for i, cog_ind in enumerate(cog_condition):
eeg_state_centers[i] = np.mean(erp_data_mean[cog_ind-1, :, ((time_index[i]+start_t)-half_win):((time_index[i]+start_t)+half_win)], axis=1)
In [ ]:
draw_states.draw_grand_average_topo(eeg_state_centers, epo, cmap='bwr', draw_separate=True, colorbar=False);
In [ ]:
draw_states.draw_topo_diff(eeg_state_centers[6,:]-eeg_state_centers[1,:], epo, title='State 7 - State 2', cmap='bwr');
draw_states.draw_topo_diff(eeg_state_centers[9,:]-eeg_state_centers[8,:], epo, title='State 10 - State 9', cmap='bwr');
# draw_states.draw_topo_diff(eeg_state_centers[7,:]-eeg_state_centers[0,:], epo, title='State 8 - State 1', cmap='bwr');
# draw_states.draw_topo_diff(eeg_state_centers[3,:]-eeg_state_centers[9,:], epo, title='State 4 - State 10', cmap='bwr');
# draw_states.draw_topo_diff(eeg_state_centers[2,:]-eeg_state_centers[3,:], epo, title='State 3 - State 4', cmap='bwr');
